window.REPORT_DATA = { "overview": { "date": "2026-05-08", "model_count": 15, "case_count": 11, "rendered_snippet_count": 322, "typescript_models": 3, "xml_models": 8, "other_dialect_models": 4, "special_models": 2, "summary": "# Tool Schema Rendering Atlas — Executive Summary\n\n_Date: 2026-05-08_\n\n## Key finding\n\nMost recent open-weight model repositories in this corpus can render some form of tool definition into model-visible prompt text, but **tool support is not one thing**. The same OpenAI-style JSON function schema may become XML tags, TypeScript declarations, a raw JSON block, DSML/custom-channel markup, or a message-embedded functions field.\n\nThe most important distinction is between:\n\n1. **Arbitrary user-defined tool schemas** supplied through normal `tools=[...]` or a custom renderer; and\n2. **Special/builtin/non-standard capabilities** such as browser search, Python execution, Jupyter/code-interpreter markers, web-tool instructions, or transcript-only tool-call formatting.\n\n## Rendering dialects at a glance\n\n| Dialect | What the model sees | Models |\n|---|---|---|\n| **XML / tag syntax** | Tool schemas wrapped in ``, calls in `` or model-specific XML-ish tags | SmolLM3, MiniMax-M2.5/M2.7, Qwen3.5/3.6, MiMo-V2.5-Pro, GLM-5/5.1 |\n| **TypeScript declarations** | `namespace functions { type tool_name = (_: { ... }) => any; }` | gpt-oss-120b, Kimi-K2.5/K2.6 |\n| **JSON schema block** | Full JSON-ish tool schema passed through in a bracketed tools block | Gemma-4-31B-it, Mistral-Medium-3.5 |\n| **Custom channel protocol** | Model-specific markers such as DeepSeek DSML `<｜DSML｜tool_calls>` | DeepSeek-V4-Pro |\n| **Message-embedded JSON** | Schema injected via message `functions` fields rather than the `tools` kwarg | Olmo-3.1-32B-Think |\n\n## Cross-model snapshot\n\n| Model | Dialect | Main support path | Renders | Special/highlight notes |\n|---|---|---|---:|---|\n| SmolLM3-3B | XML/tags | Standard `tools` arg | 10/10 | Rendered `python_tools` and `xml_tools` kwargs |\n| Qwen3.5-397B-A17B | XML/tags | Standard `tools` arg | 10/10 | Tool-call/tool-response XML transcript syntax |\n| Qwen3.6-35B-A3B | XML/tags | Standard `tools` arg | 10/10 | Tool-call/tool-response XML transcript syntax |\n| Mistral-Medium-3.5 | JSON | Standard `tools` arg | 10/10 | Web/fetch instructions require dedicated supplied tools; no builtin web tool found |\n| GLM-5 | XML/tags | Standard `tools` arg | 10/10 | Custom ``/`` call format |\n| GLM-5.1 | XML/tags | Standard `tools` arg | 10/10 | Custom ``/`` call format |\n| MiMo-V2.5-Pro | XML/tags | Standard `tools` arg | 10/10 | Transcript-format probe for `` / `` |\n| gpt-oss-120b | TypeScript | Standard `tools` arg | 9/10 | Rendered browser and Python builtins; prompt-setting probes separated |\n| MiniMax-M2.5 | XML/tags | Standard `tools` arg | 9/10 | Static `` / `` tokenizer hints only |\n| MiniMax-M2.7 | XML/tags | Standard `tools` arg | 9/10 | Static `` / `` tokenizer hints only |\n| Gemma-4-31B-it | JSON/custom tokens | Standard `tools` arg | 9/10 | Tool-call special tokens, no builtin browser/Python branch found |\n| DeepSeek-V4-Pro | Custom DSML | Custom encoder path | 9/10 | Search-agent protocol: `search`/`open`/`find` over DSML plus routing/query/read-url task tokens; schema-provided/internal path, not builtin |\n| Kimi-K2.5 | TypeScript | Custom TS renderer | 8/10 | Custom tool-call section tokens; custom renderer required |\n| Kimi-K2.6 | TypeScript | Custom TS renderer | 8/10 | Custom tool-call section tokens; custom renderer required |\n| Olmo-3.1-32B-Think | Message JSON | Message `functions` field | 10/10 | `` / `environment` message protocol; standard `tools` kwarg not meaningful |\n\n## Special, builtin, latent, and transcript-only capabilities\n\n| Model | Type | What to surface | Interpretation |\n|---|---|---|---|\n| **SmolLM3-3B** | Rendered special kwargs | `python_tools` produces ``-based stateful Jupyter/Python instructions; `xml_tools` produces XML tool calling | Strong template/render evidence for non-standard tool modes |\n| **gpt-oss-120b** | Rendered builtin tools | `builtin_tools=[\"browser\"]` renders `namespace browser` with `search`, `open`, `find`; `builtin_tools=[\"python\"]` renders stateful Python/Jupyter execution instructions | Strong template/render evidence for builtin browser and Python modes |\n| **gpt-oss-120b** | Prompt settings | `model_identity`, `reasoning_effort`, and `medium` alter prompt/settings | Useful, but not callable tools; do not count as special tool modes |\n| **MiniMax-M2.5/M2.7** | Static tokenizer hints | ``, ``, ``, ``, `` | Worth highlighting as latent/static evidence only; no active special kwarg/rendered branch found |\n| **Mistral-Medium-3.5** | Instructional web-tool note | Default prompt says to use dedicated tools for web search/fetch/up-to-date information and says it cannot browse without tools | Important user-facing policy/instruction, but not a builtin `web_search` or `web_fetch` tool |\n| **DeepSeek-V4-Pro** | Custom search-agent protocol | Encoder fixture supplies web-style tools `search`, `open`, and `find` through `developer.tools` and renders them using DSML; README also documents `<｜action｜>`, `<｜query｜>`, and `<｜read_url｜>` search-agent task tokens | Strong evidence for a model-specific web-search workflow; schema-provided/internal-search-agent path, not a public builtin `web_search`/`web_fetch` template kwarg |\n| **MiMo-V2.5-Pro** | Transcript-format support | Prior assistant `message.tool_calls` and `role='tool'` messages serialize as `` and `` | Shows how a completed/ongoing tool loop is represented in the conversation transcript, not a separate builtin capability |\n\n## What “transcript-format support” means\n\nA model template has two different jobs:\n\n1. **Declare available tools** so the next assistant turn knows what tools it may call.\n2. **Serialize the conversation history** after a tool call has already happened, including the assistant's previous tool call and the tool's response.\n\nFor MiMo, the ad hoc probe tested the second job. It supplied a conversation containing:\n\n- a user request;\n- an assistant message with `tool_calls=[...]` for `lookup_city`;\n- a following `role=\"tool\"` message containing the tool result;\n- a final assistant answer.\n\nThe template rendered those historical messages as:\n\n- ` ... ... Paris ... `\n- ` ... `\n\nThat is important because servers, agents, and fine-tuning datasets need to know the exact transcript syntax. But it does **not** mean MiMo exposes a special builtin like browser search or Python execution. It means MiMo has a tool-call/tool-response transcript format.\n\n## DeepSeek search-agent protocol note\n\n**DeepSeek-V4-Pro has a documented custom search-agent protocol.** Its encoder fixture supplies web-style tools named `search`, `open`, and `find` through a `developer.tools` field and renders them using DSML. The encoder also documents search-routing/query/read-url special tokens (`<｜action｜>`, `<｜query｜>`, `<｜read_url｜>`). This is strong evidence for a model-specific web-search workflow, but the tools are schema-provided in the conversation and the `developer` role is described as internal search-agent-only, so it should not be counted as a public builtin `web_search`/`web_fetch` template kwarg.\n\n## Schema feature survival\n\nBasic features — tool name, description, parameters, properties, required fields — survive in nearly all meaningful renders. Advanced features diverge:\n\n- **`anyOf` / `$ref` / `$defs`**: often preserved by XML/JSON dialects; often collapsed by TypeScript renderers.\n- **`oneOf` / `allOf`**: preserved by JSON-passthrough and many XML models; can collapse or error in TypeScript renderers.\n- **`additionalProperties`**: usually visible in JSON/XML paths; commonly omitted in TypeScript declarations.\n- **`default` / `minimum` / `const`**: mixed; JSON-passthrough paths tend to preserve more detail than declaration-style paths.\n\n## Limitations\n\n- **Template evidence only.** Hosted APIs or serving layers may normalize, reject, or rewrite tool schemas before the model receives them.\n- **No runtime generation tests.** A meaningful rendered prompt does not prove the model will emit correct tool calls.\n- **Direct-Jinja/custom renderer caveat.** Some custom paths approximate or invoke repo-specific rendering rather than a standard HF tokenizer path.\n- **Single snapshot.** Model revisions can update chat templates and special tokens.\n" }, "models": [ { "model": "HuggingFaceTB/SmolLM3-3B", "slug": "HuggingFaceTB__SmolLM3-3B", "revision": "a07cc9a04f16550a088caea529712d1d335b0ac1", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "xml/tag syntax", "dialect": "xml", "standard_supported": 10, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 3, "special_total": 3, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "xml/tag syntax": 10 }, "token_min": 260, "token_median": 260, "token_max": 577, "features": { "tool_name": { "positive": 10, "applicable": 20 }, "description": { "positive": 10, "applicable": 20 }, "parameters": { "positive": 9, "applicable": 18 }, "properties": { "positive": 9, "applicable": 18 }, "required": { "positive": 9, "applicable": 18 }, "object": { "positive": 9, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 4, "applicable": 4 }, "oneOf": { "positive": 2, "applicable": 2 }, "allOf": { "positive": 2, "applicable": 2 }, "ref": { "positive": 4, "applicable": 4 }, "defs": { "positive": 4, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 1, "applicable": 2 }, "additionalProperties": { "positive": 5, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 3, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/minimal_string_required.tools_argument.rendered.txt", "token_count": 404, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/minimal_string_required.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/primitive_types.tools_argument.rendered.txt", "token_count": 437, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/primitive_types.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 518, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 495, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 453, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 577, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/raw_json_schema_parameters_only.tools_argument.rendered.txt", "token_count": 400, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/function_without_parameters.tools_argument.rendered.txt", "token_count": 363, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/function_without_parameters.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 472, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 527, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "transformed_or_collapsed_to_any", "allOf": "transformed_or_collapsed_to_any", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/empty_tools.tools_argument.rendered.txt", "token_count": 260, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/empty_tools.message_functions_json.rendered.txt", "token_count": 260, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/HuggingFaceTB__SmolLM3-3B.analysis.json", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, builtin_special_tool_kwargs, python_tool_style, xml_tool_style, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/HuggingFaceTB__SmolLM3-3B.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': True, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': False, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 51 ], "snippet": "{%- set ns = namespace(xml_tool_string=\"You may call one or more functions to assist with the user query.\\nYou are provided with function signatures within XML tags:\\n\\n\\n\") -%}" }, { "file": "chat_template.jinja", "lines": [ 55 ], "snippet": "{%- set xml_tool_string = ns.xml_tool_string + \"\\n\\nFor each function call, return a json object with function name and arguments within XML tags:\\n\\n{\\\"name\\\": , \\\"arguments\\\": }\\n\" -%}" }, { "file": "chat_template.jinja", "lines": [ 59 ], "snippet": "{%- set ns = namespace(python_tool_string=\"When you send a message containing Python code between '' and '' tags, it will be executed in a stateful Jupyter notebook environment, and you will then be given the output to continued reasoning in an agentic loop.\\n\\nYou can use the following tools in your python code like regular functions:\\n\\n\") -%}" }, { "file": "README.md", "lines": [ 145 ], "snippet": "- Under the argument `xml_tools` for standard tool-calling: these tools will be called as JSON blobs within XML tags, like `{\"name\": \"get_weather\", \"arguments\": {\"city\": \"Copenhagen\"}}`" }, { "file": "README.md", "lines": [ 146 ], "snippet": "- Or under `python_tools`: then the model will call tools like python functions in a `` snippet, like `get_weather(city=\"Copenhagen\")`" }, { "file": "tokenizer_config.json", "lines": [ 124 ], "snippet": "\"content\": \"\"," }, { "file": "tokenizer_config.json", "lines": [ 132 ], "snippet": "\"content\": \"\"," }, { "file": "chat_template.jinja", "lines": [ 45 ], "snippet": "{%- if xml_tools or python_tools or tools -%}" }, { "file": "chat_template.jinja", "lines": [ 46 ], "snippet": "{{- \"### Tools\\n\\n\" -}}" } ] }, { "claim": "Rendered prompt artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'transformed_or_collapsed_to_any', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'transformed_or_collapsed_to_any', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'transformed_or_collapsed_to_any', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=404" }, { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=260" }, { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=437" }, { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=260" }, { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=518" }, { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=260" }, { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=495" }, { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=260" }, { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=453" }, { "artifact": "research/artifacts/tokenization/HuggingFaceTB__SmolLM3-3B/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=260" } ] }, { "claim": "Builtin/special/non-JSON tool-mode artifacts were produced for 3 probe(s), separate from JSON Schema `tools=[...]` support.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Special/builtin modes are template-specific kwargs and are not equivalent to arbitrary JSON Schema function tools." ], "evidence": [ { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/special_tool_modes/python_tools_minimal.rendered.txt", "observation": "probe=python_tools_minimal, style=python_tools, method=transformers_apply_chat_template:special_tool_mode, markers=['', 'python code', 'stateful Jupyter', 'regular functions'], changed_from_baseline=True" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/special_tool_modes/python_tools_openai_wrapped.rendered.txt", "observation": "probe=python_tools_openai_wrapped, style=python_tools, method=transformers_apply_chat_template:special_tool_mode, markers=['', 'python code', 'stateful Jupyter', 'calculate', 'regular functions'], changed_from_baseline=True" }, { "artifact": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/special_tool_modes/xml_tools_minimal.rendered.txt", "observation": "probe=xml_tools_minimal, style=xml_tools, method=transformers_apply_chat_template:special_tool_mode, markers=['', '', 'lookup_city', 'function signatures'], changed_from_baseline=True" } ] } ] }, "special_modes": [ { "probe_id": "python_tools_minimal", "style_family": "python_tools", "candidate_kwarg": "python_tools", "category": "builtin/rendered special tool", "counts_as_special_tool": true, "description": "Rendered evidence for a non-standard tool kwarg or builtin tool mode.", "meaningful": true, "changed_from_baseline": true, "markers_found": ", python code, stateful Jupyter, regular functions", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/special_tool_modes/python_tools_minimal.rendered.txt", "token_count": 369 }, { "probe_id": "python_tools_openai_wrapped", "style_family": "python_tools", "candidate_kwarg": "python_tools", "category": "builtin/rendered special tool", "counts_as_special_tool": true, "description": "Rendered evidence for a non-standard tool kwarg or builtin tool mode.", "meaningful": true, "changed_from_baseline": true, "markers_found": ", python code, stateful Jupyter, calculate, regular functions", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/special_tool_modes/python_tools_openai_wrapped.rendered.txt", "token_count": 416 }, { "probe_id": "xml_tools_minimal", "style_family": "xml_tools", "candidate_kwarg": "xml_tools", "category": "builtin/rendered special tool", "counts_as_special_tool": true, "description": "Rendered evidence for a non-standard tool kwarg or builtin tool mode.", "meaningful": true, "changed_from_baseline": true, "markers_found": ", , lookup_city, function signatures", "rendered_path": "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/special_tool_modes/xml_tools_minimal.rendered.txt", "token_count": 398 } ], "highlighted_notes": [] }, { "model": "MiniMaxAI/MiniMax-M2.5", "slug": "MiniMaxAI__MiniMax-M2.5", "revision": "f710177d938eff80b684d42c5aa84b382612f21f", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "xml/tag syntax", "dialect": "xml", "standard_supported": 9, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "xml/tag syntax": 9 }, "token_min": 50, "token_median": 50, "token_max": 382, "features": { "tool_name": { "positive": 9, "applicable": 20 }, "description": { "positive": 9, "applicable": 20 }, "parameters": { "positive": 8, "applicable": 18 }, "properties": { "positive": 8, "applicable": 18 }, "required": { "positive": 8, "applicable": 18 }, "object": { "positive": 8, "applicable": 18 }, "array": { "positive": 4, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 2, "applicable": 4 }, "oneOf": { "positive": 1, "applicable": 2 }, "allOf": { "positive": 1, "applicable": 2 }, "ref": { "positive": 2, "applicable": 4 }, "defs": { "positive": 2, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 1, "applicable": 2 }, "additionalProperties": { "positive": 4, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 3, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/minimal_string_required.tools_argument.rendered.txt", "token_count": 205, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/minimal_string_required.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/primitive_types.tools_argument.rendered.txt", "token_count": 238, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/primitive_types.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 319, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "present_exact_key", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 299, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 250, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 382, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "TypeError: Object of type Undefined is not JSON serializable; TypeError: Object of type StrictUndefined is not JSON serializable" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/function_without_parameters.tools_argument.rendered.txt", "token_count": 165, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/function_without_parameters.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 268, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 329, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "present_exact_key", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/empty_tools.tools_argument.rendered.txt", "token_count": 50, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/empty_tools.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/MiniMaxAI__MiniMax-M2.5.analysis.json", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, python_tool_style, xml_tool_style, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/MiniMaxAI__MiniMax-M2.5.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': False, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 67 ], "snippet": "{{- render_tool_namespace(\"functions\", tools) }}" }, { "file": "chat_template.jinja", "lines": [ 81 ], "snippet": "{%- set last_tool_call = namespace(name=none) -%}" }, { "file": "chat_template.jinja", "lines": [ 107 ], "snippet": "{%- if tool_call.function %}" }, { "file": "chat_template.jinja", "lines": [ 108 ], "snippet": "{%- set tool_call = tool_call.function %}" }, { "file": "chat_template.jinja", "lines": [ 2 ], "snippet": "{%- set toolcall_begin_token = '' -%}" }, { "file": "chat_template.jinja", "lines": [ 3 ], "snippet": "{%- set toolcall_end_token = '' -%}" }, { "file": "chat_template.jinja", "lines": [ 4 ], "snippet": "{#- Tool Rendering Functions ============================================== -#}" }, { "file": "chat_template.jinja", "lines": [ 5 ], "snippet": "{%- macro render_tool_namespace(namespace_name, tool_list) -%}" }, { "file": "chat_template.jinja", "lines": [ 7 ], "snippet": "{{ tool.function | tojson(ensure_ascii=False) }}" } ] }, { "claim": "Rendered prompt artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=205" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=50" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=238" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=50" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=319" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=50" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=299" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=50" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=250" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.5/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=50" } ] } ] }, "special_modes": [], "highlighted_notes": [ { "category": "static token hint", "title": "Latent code/Jupyter tokens", "detail": "Tokenizer config contains `` and `///` tokens. The inspected chat template does not expose a `python_tools`, `builtin_tools`, or `code_interpreter` kwarg branch, so this is a static/tokenizer hint rather than confirmed Python execution support.", "evidence": "research/artifacts/hf-snapshots/MiniMaxAI__MiniMax-M2.5/tokenizer_config.json" } ] }, { "model": "MiniMaxAI/MiniMax-M2.7", "slug": "MiniMaxAI__MiniMax-M2.7", "revision": "d494266a4affc0d2995ba1fa35c8481cbd84294b", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "xml/tag syntax", "dialect": "xml", "standard_supported": 9, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "xml/tag syntax": 9 }, "token_min": 50, "token_median": 50, "token_max": 382, "features": { "tool_name": { "positive": 9, "applicable": 20 }, "description": { "positive": 9, "applicable": 20 }, "parameters": { "positive": 8, "applicable": 18 }, "properties": { "positive": 8, "applicable": 18 }, "required": { "positive": 8, "applicable": 18 }, "object": { "positive": 8, "applicable": 18 }, "array": { "positive": 4, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 2, "applicable": 4 }, "oneOf": { "positive": 1, "applicable": 2 }, "allOf": { "positive": 1, "applicable": 2 }, "ref": { "positive": 2, "applicable": 4 }, "defs": { "positive": 2, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 1, "applicable": 2 }, "additionalProperties": { "positive": 4, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 3, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/minimal_string_required.tools_argument.rendered.txt", "token_count": 205, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/minimal_string_required.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/primitive_types.tools_argument.rendered.txt", "token_count": 238, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/primitive_types.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 319, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "present_exact_key", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 299, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 250, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 382, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "TypeError: Object of type Undefined is not JSON serializable; TypeError: Object of type StrictUndefined is not JSON serializable" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/function_without_parameters.tools_argument.rendered.txt", "token_count": 165, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/function_without_parameters.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 268, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 329, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "present_exact_key", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/empty_tools.tools_argument.rendered.txt", "token_count": 50, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/empty_tools.message_functions_json.rendered.txt", "token_count": 50, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/MiniMaxAI__MiniMax-M2.7.analysis.json", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, python_tool_style, xml_tool_style, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/MiniMaxAI__MiniMax-M2.7.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': False, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 67 ], "snippet": "{{- render_tool_namespace(\"functions\", tools) }}" }, { "file": "chat_template.jinja", "lines": [ 81 ], "snippet": "{%- set last_tool_call = namespace(name=none) -%}" }, { "file": "chat_template.jinja", "lines": [ 107 ], "snippet": "{%- if tool_call.function %}" }, { "file": "chat_template.jinja", "lines": [ 108 ], "snippet": "{%- set tool_call = tool_call.function %}" }, { "file": "chat_template.jinja", "lines": [ 2 ], "snippet": "{%- set toolcall_begin_token = '' -%}" }, { "file": "chat_template.jinja", "lines": [ 3 ], "snippet": "{%- set toolcall_end_token = '' -%}" }, { "file": "chat_template.jinja", "lines": [ 4 ], "snippet": "{#- Tool Rendering Functions ============================================== -#}" }, { "file": "chat_template.jinja", "lines": [ 5 ], "snippet": "{%- macro render_tool_namespace(namespace_name, tool_list) -%}" }, { "file": "chat_template.jinja", "lines": [ 7 ], "snippet": "{{ tool.function | tojson(ensure_ascii=False) }}" } ] }, { "claim": "Rendered prompt artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=205" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=50" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=238" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=50" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=319" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=50" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=299" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=50" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=250" }, { "artifact": "research/artifacts/tokenization/MiniMaxAI__MiniMax-M2.7/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=50" } ] } ] }, "special_modes": [], "highlighted_notes": [ { "category": "static token hint", "title": "Latent code/Jupyter tokens", "detail": "Tokenizer config contains `` and `///` tokens. The inspected chat template does not expose a `python_tools`, `builtin_tools`, or `code_interpreter` kwarg branch, so this is a static/tokenizer hint rather than confirmed Python execution support.", "evidence": "research/artifacts/hf-snapshots/MiniMaxAI__MiniMax-M2.7/tokenizer_config.json" } ] }, { "model": "Qwen/Qwen3.5-397B-A17B", "slug": "Qwen__Qwen3.5-397B-A17B", "revision": "8472618112abcbd45acbcdc58436aff4233c23f7", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "xml/tag syntax", "dialect": "xml", "standard_supported": 10, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "xml/tag syntax": 10 }, "token_min": 22, "token_median": 22, "token_max": 460, "features": { "tool_name": { "positive": 10, "applicable": 20 }, "description": { "positive": 10, "applicable": 20 }, "parameters": { "positive": 9, "applicable": 18 }, "properties": { "positive": 9, "applicable": 18 }, "required": { "positive": 9, "applicable": 18 }, "object": { "positive": 9, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 2, "applicable": 4 }, "oneOf": { "positive": 1, "applicable": 2 }, "allOf": { "positive": 1, "applicable": 2 }, "ref": { "positive": 2, "applicable": 4 }, "defs": { "positive": 2, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 1, "applicable": 2 }, "additionalProperties": { "positive": 5, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 3, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/minimal_string_required.tools_argument.rendered.txt", "token_count": 287, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/minimal_string_required.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/primitive_types.tools_argument.rendered.txt", "token_count": 320, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/primitive_types.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 401, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 378, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 336, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 460, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/raw_json_schema_parameters_only.tools_argument.rendered.txt", "token_count": 284, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/function_without_parameters.tools_argument.rendered.txt", "token_count": 247, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/function_without_parameters.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 353, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 410, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/empty_tools.tools_argument.rendered.txt", "token_count": 22, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/empty_tools.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/Qwen__Qwen3.5-397B-A17B.analysis.json", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, has_tokenizer_config_chat_template, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, xml_tool_style, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/Qwen__Qwen3.5-397B-A17B.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': True, 'tokenizer_config_chat_template_support': True, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 53 ], "snippet": "{{- '\\n\\nIf you choose to call a function ONLY reply in the following format with NO suffix:\\n\\n\\n\\n\\nvalue_1\\n\\n\\nThis is the value for the second parameter\\nthat can span\\nmultiple lines\\n\\n\\n\\n\\n\\nReminder:\\n- Function calls MUST follow the specified format: an inner block must be nested within <" }, { "file": "chat_template.jinja", "lines": [ 107 ], "snippet": "{%- if tool_call.function is defined %}" }, { "file": "chat_template.jinja", "lines": [ 108 ], "snippet": "{%- set tool_call = tool_call.function %}" }, { "file": "chat_template.jinja", "lines": [ 112 ], "snippet": "{{- '\\n\\n\\n\\n' }}" }, { "file": "chat_template.jinja", "lines": [ 114 ], "snippet": "{{- '\\n\\n' }}" }, { "file": "chat_template.jinja", "lines": [ 117 ], "snippet": "{{- '\\n\\n\\n' }}" }, { "file": "chat_template.jinja", "lines": [ 127 ], "snippet": "{{- '\\n' }}" }, { "file": "chat_template.jinja", "lines": [ 47 ], "snippet": "{{- \"# Tools\\n\\nYou have access to the following functions:\\n\\n\" }}" }, { "file": "chat_template.jinja", "lines": [ 67 ], "snippet": "{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}" } ] }, { "claim": "Rendered prompt artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=287" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=22" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=320" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=22" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=401" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=22" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=378" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=22" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=336" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.5-397B-A17B/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=22" } ] } ] }, "special_modes": [], "highlighted_notes": [] }, { "model": "Qwen/Qwen3.6-35B-A3B", "slug": "Qwen__Qwen3.6-35B-A3B", "revision": "995ad96eacd98c81ed38be0c5b274b04031597b0", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "xml/tag syntax", "dialect": "xml", "standard_supported": 10, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "xml/tag syntax": 10 }, "token_min": 22, "token_median": 22, "token_max": 460, "features": { "tool_name": { "positive": 10, "applicable": 20 }, "description": { "positive": 10, "applicable": 20 }, "parameters": { "positive": 9, "applicable": 18 }, "properties": { "positive": 9, "applicable": 18 }, "required": { "positive": 9, "applicable": 18 }, "object": { "positive": 9, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 2, "applicable": 4 }, "oneOf": { "positive": 1, "applicable": 2 }, "allOf": { "positive": 1, "applicable": 2 }, "ref": { "positive": 2, "applicable": 4 }, "defs": { "positive": 2, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 1, "applicable": 2 }, "additionalProperties": { "positive": 5, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 3, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/minimal_string_required.tools_argument.rendered.txt", "token_count": 287, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/minimal_string_required.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/primitive_types.tools_argument.rendered.txt", "token_count": 320, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/primitive_types.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 401, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 378, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 336, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 460, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/raw_json_schema_parameters_only.tools_argument.rendered.txt", "token_count": 284, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/function_without_parameters.tools_argument.rendered.txt", "token_count": 247, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/function_without_parameters.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 353, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 410, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/empty_tools.tools_argument.rendered.txt", "token_count": 22, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/empty_tools.message_functions_json.rendered.txt", "token_count": 22, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/Qwen__Qwen3.6-35B-A3B.analysis.json", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, has_tokenizer_config_chat_template, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, xml_tool_style, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/Qwen__Qwen3.6-35B-A3B.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': True, 'tokenizer_config_chat_template_support': True, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 53 ], "snippet": "{{- '\\n\\nIf you choose to call a function ONLY reply in the following format with NO suffix:\\n\\n\\n\\n\\nvalue_1\\n\\n\\nThis is the value for the second parameter\\nthat can span\\nmultiple lines\\n\\n\\n\\n\\n\\nReminder:\\n- Function calls MUST follow the specified format: an inner block must be nested within <" }, { "file": "chat_template.jinja", "lines": [ 107 ], "snippet": "{%- if tool_call.function is defined %}" }, { "file": "chat_template.jinja", "lines": [ 108 ], "snippet": "{%- set tool_call = tool_call.function %}" }, { "file": "chat_template.jinja", "lines": [ 112 ], "snippet": "{{- '\\n\\n\\n\\n' }}" }, { "file": "chat_template.jinja", "lines": [ 114 ], "snippet": "{{- '\\n\\n' }}" }, { "file": "chat_template.jinja", "lines": [ 117 ], "snippet": "{{- '\\n\\n\\n' }}" }, { "file": "chat_template.jinja", "lines": [ 127 ], "snippet": "{{- '\\n' }}" }, { "file": "chat_template.jinja", "lines": [ 47 ], "snippet": "{{- \"# Tools\\n\\nYou have access to the following functions:\\n\\n\" }}" }, { "file": "chat_template.jinja", "lines": [ 67 ], "snippet": "{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}" } ] }, { "claim": "Rendered prompt artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=287" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=22" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=320" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=22" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=401" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=22" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=378" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=22" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=336" }, { "artifact": "research/artifacts/tokenization/Qwen__Qwen3.6-35B-A3B/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=22" } ] } ] }, "special_modes": [], "highlighted_notes": [] }, { "model": "XiaomiMiMo/MiMo-V2.5-Pro", "slug": "XiaomiMiMo__MiMo-V2.5-Pro", "revision": "f5953f1b993e90367f3d87fed59d099f2c527763", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "xml/tag syntax", "dialect": "xml", "standard_supported": 10, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 1, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "xml/tag syntax": 10 }, "token_min": 263, "token_median": 263, "token_max": 714, "features": { "tool_name": { "positive": 10, "applicable": 20 }, "description": { "positive": 10, "applicable": 20 }, "parameters": { "positive": 18, "applicable": 18 }, "properties": { "positive": 6, "applicable": 18 }, "required": { "positive": 9, "applicable": 18 }, "object": { "positive": 4, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 4, "applicable": 4 }, "oneOf": { "positive": 2, "applicable": 2 }, "allOf": { "positive": 2, "applicable": 2 }, "ref": { "positive": 4, "applicable": 4 }, "defs": { "positive": 4, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 2, "applicable": 2 }, "additionalProperties": { "positive": 5, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 3, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/minimal_string_required.tools_argument.rendered.txt", "token_count": 498, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "omitted", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/minimal_string_required.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "present_exact_key", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/primitive_types.tools_argument.rendered.txt", "token_count": 563, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "omitted", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/primitive_types.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "present_exact_key", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 639, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "present_exact_key", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 626, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "present_exact_key", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 550, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "omitted", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "present_exact_key", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 714, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "present_exact_key", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/raw_json_schema_parameters_only.tools_argument.rendered.txt", "token_count": 508, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "present_exact_key", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/function_without_parameters.tools_argument.rendered.txt", "token_count": 456, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/function_without_parameters.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 585, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "present_exact_key", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 656, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "present_exact_key", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "transformed_or_collapsed_to_any", "allOf": "transformed_or_collapsed_to_any", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/empty_tools.tools_argument.rendered.txt", "token_count": 263, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/empty_tools.message_functions_json.rendered.txt", "token_count": 263, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/XiaomiMiMo__MiMo-V2.5-Pro.analysis.json", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_tokenizer_config_chat_template, custom_tool_tags_or_channels, xml_tool_style, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/XiaomiMiMo__MiMo-V2.5-Pro.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': False, 'non_json_tool_style_hints': True, 'processor_static_only': False, 'tokenizer_config_chat_template_support': True, 'tool_named_code_or_template': False}" }, { "file": "added_tokens.json", "lines": [ 3 ], "snippet": "\"\": 151658," }, { "file": "added_tokens.json", "lines": [ 6 ], "snippet": "\"\": 151657," }, { "file": "tokenizer_config.json", "lines": [ 118 ], "snippet": "\"content\": \"\"," }, { "file": "tokenizer_config.json", "lines": [ 126 ], "snippet": "\"content\": \"\"," }, { "file": "modeling_mimo_v2.py", "lines": [ 23 ], "snippet": "import torch.nn.functional as F" }, { "file": "modeling_mimo_v2.py", "lines": [ 31 ], "snippet": "from transformers.modeling_rope_utils import ROPE_INIT_FUNCTIONS, dynamic_rope_update" }, { "file": "modeling_mimo_v2.py", "lines": [ 32 ], "snippet": "from transformers.modeling_utils import ALL_ATTENTION_FUNCTIONS, PreTrainedModel" }, { "file": "modeling_mimo_v2.py", "lines": [ 94 ], "snippet": "attn_weights = nn.functional.dropout(probs, p=dropout, training=module.training)" }, { "file": "modeling_mimo_v2.py", "lines": [ 154 ], "snippet": "raise NotImplementedError(f\"Unsupported scoring function for MoE gating: {self.scoring_func}\")" } ] }, { "claim": "Rendered prompt artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'transformed_or_collapsed_to_any', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'omitted', 'defs': 'transformed_or_collapsed_to_any', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'omitted', 'ref': 'transformed_or_collapsed_to_any', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=498" }, { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=263" }, { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=563" }, { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=263" }, { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=639" }, { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=263" }, { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=626" }, { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=263" }, { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=550" }, { "artifact": "research/artifacts/tokenization/XiaomiMiMo__MiMo-V2.5-Pro/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=263" } ] } ] }, "special_modes": [ { "probe_id": "ad_hoc_mimo_tool_messages", "style_family": "transcript-format support", "candidate_kwarg": "ad_hoc_message_probe", "category": "transcript-format support", "counts_as_special_tool": false, "description": "Shows serialization of prior assistant tool calls and tool responses; not a separate builtin tool.", "meaningful": true, "changed_from_baseline": true, "markers_found": "", "rendered_path": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/special_tool_modes/ad_hoc_mimo_tool_messages.rendered.txt", "token_count": 319 } ], "highlighted_notes": [ { "category": "transcript-format support", "title": "Tool-call / tool-response transcript format", "detail": "An ad hoc probe shows how prior assistant `message.tool_calls` and subsequent `role='tool'` messages are serialized as `` and `` blocks. This is transcript formatting for an already-occurring tool loop, not a separate builtin capability like browser or Python execution.", "evidence": "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/special_tool_modes/ad_hoc_mimo_tool_messages.probe.md" } ] }, { "model": "allenai/Olmo-3.1-32B-Think", "slug": "allenai__Olmo-3.1-32B-Think", "revision": "832c3f543499af8fe68b88359501de9cb7840544", "support_variant": "message_functions_json", "support_path": "Message functions JSON", "support_detail": "Schema evidence appears mainly through message-embedded function JSON.", "primary_style": "message functions json", "dialect": "message-json", "standard_supported": 0, "message_supported": 10, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "message functions json": 10 }, "token_min": 68, "token_median": 74, "token_max": 408, "features": { "tool_name": { "positive": 10, "applicable": 20 }, "description": { "positive": 10, "applicable": 20 }, "parameters": { "positive": 9, "applicable": 18 }, "properties": { "positive": 9, "applicable": 18 }, "required": { "positive": 9, "applicable": 18 }, "object": { "positive": 9, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 2, "applicable": 4 }, "oneOf": { "positive": 1, "applicable": 2 }, "allOf": { "positive": 1, "applicable": 2 }, "ref": { "positive": 2, "applicable": 4 }, "defs": { "positive": 2, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 1, "applicable": 2 }, "additionalProperties": { "positive": 5, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 3, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/minimal_string_required.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/minimal_string_required.message_functions_json.rendered.txt", "token_count": 166, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/primitive_types.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/primitive_types.message_functions_json.rendered.txt", "token_count": 215, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 346, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 313, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 241, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 408, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/raw_json_schema_parameters_only.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 161, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/function_without_parameters.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/function_without_parameters.message_functions_json.rendered.txt", "token_count": 107, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 251, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" }, "message_functions_json": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 359, "style": "message functions json", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/empty_tools.tools_argument.rendered.txt", "token_count": 68, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/empty_tools.message_functions_json.rendered.txt", "token_count": 74, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/minimal_string_required.message_functions_json.rendered.txt", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nested_arrays_enums.message_functions_json.rendered.txt", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nullable_anyof_refs.message_functions_json.rendered.txt", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/oneof_allof_constraints.message_functions_json.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/allenai__Olmo-3.1-32B-Think.analysis.json", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, has_tokenizer_config_chat_template, custom_tool_tags_or_channels, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/allenai__Olmo-3.1-32B-Think.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': False, 'jinja_template_tools_branch': False, 'non_json_tool_style_hints': True, 'processor_static_only': False, 'tokenizer_config_chat_template_support': True, 'tool_named_code_or_template': False}" }, { "file": "tokenizer_config.json", "lines": [ 189 ], "snippet": "\"chat_template\": \"{% set has_system = messages|selectattr('role', 'equalto', 'system')|list|length > 0 %}{% if not has_system %}{{ '<|im_start|>system\\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\\n' }}{% endif %}{% for message in messages %}{% if message['role'] == 'system' %}{{ '<|im_start|>system\\n' + message['content'] }}{% if message.get('functions', none) is not none %" }, { "file": "chat_template.jinja", "lines": [ 4 ], "snippet": "' + message['content'] }}{% if message.get('functions', none) is not none %}{{ ' ' + message['functions'] + '<|im_end|>" }, { "file": "chat_template.jinja", "lines": [ 5 ], "snippet": "' }}{% else %}{{ ' You do not currently have access to any functions. <|im_end|>" }, { "file": "chat_template.jinja", "lines": [ 6 ], "snippet": "' }}{% endif %}{% elif message['role'] == 'user' %}{% if message.get('functions', none) is not none %}{{ '<|im_start|>user" }, { "file": "chat_template.jinja", "lines": [ 8 ], "snippet": "' + '' + message['functions'] + '<|im_end|>" }, { "file": "chat_template.jinja", "lines": [ 12 ], "snippet": "' }}{% if message.get('content', none) is not none %}{{ message['content'] }}{% endif %}{% if message.get('function_calls', none) is not none %}{{ '' + message['function_calls'] + '' }}{% endif %}{% if not loop.last %}{{ '<|im_end|>' + '" }, { "file": "model.safetensors.index.json", "lines": [ 3 ], "snippet": "\"total_parameters\": 32233522176," }, { "file": "vocab.json", "lines": [ 1 ], "snippet": "{\"!\":0,\"\\\"\":1,\"#\":2,\"$\":3,\"%\":4,\"&\":5,\"'\":6,\"(\":7,\")\":8,\"*\":9,\"+\":10,\",\":11,\"-\":12,\".\":13,\"/\":14,\"0\":15,\"1\":16,\"2\":17,\"3\":18,\"4\":19,\"5\":20,\"6\":21,\"7\":22,\"8\":23,\"9\":24,\":\":25,\";\":26,\"<\":27,\"=\":28,\">\":29,\"?\":30,\"@\":31,\"A\":32,\"B\":33,\"C\":34,\"D\":35,\"E\":36,\"F\":37,\"G\":38,\"H\":39,\"I\":40,\"J\":41,\"K\":42,\"L\":43,\"M\":44,\"N\":45,\"O\":46,\"P\":47,\"Q\":48,\"R\":49,\"S\":50,\"T\":51,\"U\":52,\"V\":53,\"W\":54,\"X\":55,\"Y\":56,\"Z\":57,\"[\":58,\"\\\\\":59,\"]\":60,\"^\":61,\"_\":62,\"`\":63,\"a\":64,\"b\":65,\"c\":66,\"d\":67,\"e\":68,\"f\":69,\"g\":70,\"h\":71,\"i\"" } ] }, { "claim": "Rendered prompt artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" } ] }, { "claim": "Tokenizer artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=68" }, { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=166" }, { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=68" }, { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=215" }, { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=68" }, { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=346" }, { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=68" }, { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=313" }, { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=68" }, { "artifact": "research/artifacts/tokenization/allenai__Olmo-3.1-32B-Think/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=241" } ] } ] }, "special_modes": [], "highlighted_notes": [] }, { "model": "deepseek-ai/DeepSeek-V4-Pro", "slug": "deepseek-ai__DeepSeek-V4-Pro", "revision": "89d501aed998d33fa4f4702102ec1bb2331e10f6", "support_variant": "custom_deepseek_dsv4", "support_path": "Custom renderer path", "support_detail": "Tool support is visible through a model-specific renderer/protocol path.", "primary_style": "custom tags/channels", "dialect": "custom-channel", "standard_supported": 0, "message_supported": 0, "custom_supported": 9, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11, "custom_deepseek_dsv4": 11 }, "style_counts": { "custom tags/channels": 9 }, "token_min": 22, "token_median": 370, "token_max": 482, "features": { "tool_name": { "positive": 9, "applicable": 30 }, "description": { "positive": 9, "applicable": 30 }, "parameters": { "positive": 8, "applicable": 27 }, "properties": { "positive": 8, "applicable": 27 }, "required": { "positive": 8, "applicable": 27 }, "object": { "positive": 8, "applicable": 27 }, "array": { "positive": 2, "applicable": 6 }, "enum": { "positive": 3, "applicable": 9 }, "anyOf": { "positive": 2, "applicable": 6 }, "oneOf": { "positive": 1, "applicable": 3 }, "allOf": { "positive": 1, "applicable": 3 }, "ref": { "positive": 2, "applicable": 6 }, "defs": { "positive": 2, "applicable": 6 }, "default": { "positive": 2, "applicable": 6 }, "const": { "positive": 1, "applicable": 3 }, "additionalProperties": { "positive": 4, "applicable": 15 }, "minimum": { "positive": 3, "applicable": 9 }, "null": { "positive": 3, "applicable": 9 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/minimal_string_required.custom_deepseek_dsv4.rendered.txt", "token_count": 299, "style": "custom tags/channels", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/primitive_types.custom_deepseek_dsv4.rendered.txt", "token_count": 329, "style": "custom tags/channels", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/nested_arrays_enums.custom_deepseek_dsv4.rendered.txt", "token_count": 412, "style": "custom tags/channels", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/nullable_anyof_refs.custom_deepseek_dsv4.rendered.txt", "token_count": 394, "style": "custom tags/channels", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/multiple_tools_parallel_candidate.custom_deepseek_dsv4.rendered.txt", "token_count": 337, "style": "custom tags/channels", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/refs_unions_defaults_constraints.custom_deepseek_dsv4.rendered.txt", "token_count": 482, "style": "custom tags/channels", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "custom tags/channels", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "KeyError: 'function'" } }, "function_without_parameters": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/function_without_parameters.custom_deepseek_dsv4.rendered.txt", "token_count": 258, "style": "custom tags/channels", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/non_ascii_long_descriptions.custom_deepseek_dsv4.rendered.txt", "token_count": 370, "style": "custom tags/channels", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_rendered", "allOf": "not_rendered", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_rendered", "allOf": "not_rendered", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/oneof_allof_constraints.custom_deepseek_dsv4.rendered.txt", "token_count": 429, "style": "custom tags/channels", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": null, "token_count": null, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "ValueError: Cannot use chat template functions because tokenizer.chat_template is not set and no template argument was passed! For information about writing templates and setting the tokenizer.chat_template attribute, please see the documentation at https://huggingface.co/docs/transformers/main/en/chat_templating" }, "custom_deepseek_dsv4": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/empty_tools.custom_deepseek_dsv4.rendered.txt", "token_count": 22, "style": "custom tags/channels", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/minimal_string_required.custom_deepseek_dsv4.rendered.txt", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/nested_arrays_enums.custom_deepseek_dsv4.rendered.txt", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/nullable_anyof_refs.custom_deepseek_dsv4.rendered.txt", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/oneof_allof_constraints.custom_deepseek_dsv4.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/deepseek-ai__DeepSeek-V4-Pro.analysis.json", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_custom_encoding_code, openai_style_tool_mentions, custom_tool_tags_or_channels, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_code", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/deepseek-ai__DeepSeek-V4-Pro.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': True, 'documented_readme_support': True, 'jinja_template_tools_branch': False, 'non_json_tool_style_hints': True, 'processor_static_only': False, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': False}" }, { "file": "encoding/encoding_dsv4.py", "lines": [ 118 ], "snippet": "\"name\": tool_call[\"function\"][\"name\"]," }, { "file": "encoding/encoding_dsv4.py", "lines": [ 119 ], "snippet": "\"arguments\": tool_call[\"function\"][\"arguments\"]," }, { "file": "encoding/test_encoding_dsv4.py", "lines": [ 33 ], "snippet": "assert parsed_tc[\"tool_calls\"][0][\"function\"][\"name\"] == \"get_weather\"" }, { "file": "encoding/test_encoding_dsv4.py", "lines": [ 34 ], "snippet": "assert json.loads(parsed_tc[\"tool_calls\"][0][\"function\"][\"arguments\"]) == {\"location\": \"Beijing\", \"unit\": \"celsius\"}" }, { "file": "encoding/encoding_dsv4.py", "lines": [ 45 ], "snippet": "assistant_msg_template: str = \"{reasoning}{content}{tool_calls}\" + eos_token" }, { "file": "encoding/encoding_dsv4.py", "lines": [ 46 ], "snippet": "assistant_msg_wo_eos_template: str = \"{reasoning}{content}{tool_calls}\"" }, { "file": "encoding/encoding_dsv4.py", "lines": [ 52 ], "snippet": "tool_call_template: str = (" }, { "file": "encoding/encoding_dsv4.py", "lines": [ 55 ], "snippet": "tool_calls_template = (" }, { "file": "encoding/encoding_dsv4.py", "lines": [ 56 ], "snippet": "\"<{dsml_token}{tc_block_name}>\\n{tool_calls}\\n\"" } ] }, { "claim": "Rendered prompt artifacts were produced for 10 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/minimal_string_required.custom_deepseek_dsv4.rendered.txt", "observation": "case=minimal_string_required, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/primitive_types.custom_deepseek_dsv4.rendered.txt", "observation": "case=primitive_types, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/nested_arrays_enums.custom_deepseek_dsv4.rendered.txt", "observation": "case=nested_arrays_enums, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/nullable_anyof_refs.custom_deepseek_dsv4.rendered.txt", "observation": "case=nullable_anyof_refs, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/multiple_tools_parallel_candidate.custom_deepseek_dsv4.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/refs_unions_defaults_constraints.custom_deepseek_dsv4.rendered.txt", "observation": "case=refs_unions_defaults_constraints, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/function_without_parameters.custom_deepseek_dsv4.rendered.txt", "observation": "case=function_without_parameters, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'not_applicable', 'oneOf': 'not_applicable', 'parameters_key': 'not_applicable', 'properties_key': 'not_applicable', 'ref': 'not_applicable', 'required_key': 'not_applicable', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'not_applicable'}" }, { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/non_ascii_long_descriptions.custom_deepseek_dsv4.rendered.txt", "observation": "case=non_ascii_long_descriptions, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/oneof_allof_constraints.custom_deepseek_dsv4.rendered.txt", "observation": "case=oneof_allof_constraints, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'present_exact_key', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'present_exact_key', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'present_exact_key', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'present_exact_key', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/empty_tools.custom_deepseek_dsv4.rendered.txt", "observation": "case=empty_tools, method=custom_deepseek_dsv4:custom_deepseek_dsv4, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'not_applicable', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'not_applicable', 'oneOf': 'not_applicable', 'parameters_key': 'not_applicable', 'properties_key': 'not_applicable', 'ref': 'not_applicable', 'required_key': 'not_applicable', 'tool_name': 'not_applicable', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'not_applicable'}" } ] }, { "claim": "Tokenizer artifacts were produced for 10 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/minimal_string_required.custom_deepseek_dsv4.tokens.json", "observation": "case=minimal_string_required, tokens=299" }, { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/primitive_types.custom_deepseek_dsv4.tokens.json", "observation": "case=primitive_types, tokens=329" }, { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/nested_arrays_enums.custom_deepseek_dsv4.tokens.json", "observation": "case=nested_arrays_enums, tokens=412" }, { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/nullable_anyof_refs.custom_deepseek_dsv4.tokens.json", "observation": "case=nullable_anyof_refs, tokens=394" }, { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/multiple_tools_parallel_candidate.custom_deepseek_dsv4.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=337" }, { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/refs_unions_defaults_constraints.custom_deepseek_dsv4.tokens.json", "observation": "case=refs_unions_defaults_constraints, tokens=482" }, { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/function_without_parameters.custom_deepseek_dsv4.tokens.json", "observation": "case=function_without_parameters, tokens=258" }, { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/non_ascii_long_descriptions.custom_deepseek_dsv4.tokens.json", "observation": "case=non_ascii_long_descriptions, tokens=370" }, { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/oneof_allof_constraints.custom_deepseek_dsv4.tokens.json", "observation": "case=oneof_allof_constraints, tokens=429" }, { "artifact": "research/artifacts/tokenization/deepseek-ai__DeepSeek-V4-Pro/empty_tools.custom_deepseek_dsv4.tokens.json", "observation": "case=empty_tools, tokens=22" } ] } ] }, "special_modes": [], "highlighted_notes": [ { "category": "custom protocol example", "title": "Search/open/find tools over DSML + search-agent task tokens", "detail": "The custom encoder test fixture defines developer-provided web tools named `search`, `open`, and `find` on a `developer` message and renders them through DeepSeek's DSML protocol. The encoder README says `developer` is used exclusively in the internal search-agent pipeline, not general API chat. Tool results are folded back as `` blocks inside user messages. The same encoder also documents quick-instruction tokens such as `<｜action｜>` for deciding Search vs Answer, `<｜query｜>` for query generation, and `<｜read_url｜>` for URL-fetch/read routing. This is strong evidence of a search-agent protocol, but not a builtin `web_search`/`web_fetch` kwarg exposed by a normal HF chat template.", "evidence": "research/artifacts/hf-snapshots/deepseek-ai__DeepSeek-V4-Pro/encoding/tests/test_input_3.json" } ] }, { "model": "google/gemma-4-31B-it", "slug": "google__gemma-4-31B-it", "revision": "ba74f5b6c647c0911554e50278d6f6f4477f9010", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "json schema block", "dialect": "json", "standard_supported": 9, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "json schema block": 9 }, "token_min": 25, "token_median": 25, "token_max": 184, "features": { "tool_name": { "positive": 9, "applicable": 20 }, "description": { "positive": 9, "applicable": 20 }, "parameters": { "positive": 8, "applicable": 18 }, "properties": { "positive": 8, "applicable": 18 }, "required": { "positive": 8, "applicable": 18 }, "object": { "positive": 8, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 0, "applicable": 4 }, "oneOf": { "positive": 0, "applicable": 2 }, "allOf": { "positive": 0, "applicable": 2 }, "ref": { "positive": 0, "applicable": 4 }, "defs": { "positive": 0, "applicable": 4 }, "default": { "positive": 0, "applicable": 4 }, "const": { "positive": 0, "applicable": 2 }, "additionalProperties": { "positive": 2, "applicable": 10 }, "minimum": { "positive": 0, "applicable": 6 }, "null": { "positive": 1, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/minimal_string_required.tools_argument.rendered.txt", "token_count": 82, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/minimal_string_required.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/primitive_types.tools_argument.rendered.txt", "token_count": 107, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/primitive_types.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 184, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 106, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 117, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 121, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "UndefinedError: 'dict object' has no attribute 'function'; UndefinedError: 'dict object' has no attribute 'function'" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/function_without_parameters.tools_argument.rendered.txt", "token_count": 44, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/function_without_parameters.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 145, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 104, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/empty_tools.tools_argument.rendered.txt", "token_count": 25, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/google__gemma-4-31B-it/empty_tools.message_functions_json.rendered.txt", "token_count": 25, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/google__gemma-4-31B-it/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/google__gemma-4-31B-it/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/google__gemma-4-31B-it/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/google__gemma-4-31B-it/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/google__gemma-4-31B-it.analysis.json", "research/artifacts/rendered/google__gemma-4-31B-it/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/google__gemma-4-31B-it.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': True, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 86 ], "snippet": "{%- macro format_function_declaration(tool_data) -%}" }, { "file": "chat_template.jinja", "lines": [ 87 ], "snippet": "declaration:{{- tool_data['function']['name'] -}}{description:<|\"|>{{- tool_data['function']['description'] -}}<|\"|>" }, { "file": "chat_template.jinja", "lines": [ 88 ], "snippet": "{%- set params = tool_data['function']['parameters'] -%}" }, { "file": "chat_template.jinja", "lines": [ 106 ], "snippet": "{%- if 'response' in tool_data['function'] -%}" }, { "file": "chat_template.jinja", "lines": [ 107 ], "snippet": "{%- set response_declaration = tool_data['function']['response'] -%}" }, { "file": "chat_template.jinja", "lines": [ 199 ], "snippet": "{{- format_function_declaration(tool) | trim -}}" }, { "file": "chat_template.jinja", "lines": [ 239 ], "snippet": "{%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}" }, { "file": "chat_template.jinja", "lines": [ 243 ], "snippet": "{%- if message['tool_calls'] -%}" }, { "file": "chat_template.jinja", "lines": [ 244 ], "snippet": "{%- for tool_call in message['tool_calls'] -%}" } ] }, { "claim": "Rendered prompt artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'omitted', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/google__gemma-4-31B-it/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=82" }, { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=25" }, { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=107" }, { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=25" }, { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=184" }, { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=25" }, { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=106" }, { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=25" }, { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=117" }, { "artifact": "research/artifacts/tokenization/google__gemma-4-31B-it/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=25" } ] } ] }, "special_modes": [], "highlighted_notes": [] }, { "model": "mistralai/Mistral-Medium-3.5-128B", "slug": "mistralai__Mistral-Medium-3.5-128B", "revision": "ed0c85631cf20929c29bc36d7cb37519ed0cdd28", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "json schema block", "dialect": "json", "standard_supported": 10, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "json schema block": 10 }, "token_min": 375, "token_median": 375, "token_max": 625, "features": { "tool_name": { "positive": 10, "applicable": 20 }, "description": { "positive": 10, "applicable": 20 }, "parameters": { "positive": 9, "applicable": 18 }, "properties": { "positive": 9, "applicable": 18 }, "required": { "positive": 9, "applicable": 18 }, "object": { "positive": 9, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 4, "applicable": 4 }, "oneOf": { "positive": 2, "applicable": 2 }, "allOf": { "positive": 2, "applicable": 2 }, "ref": { "positive": 4, "applicable": 4 }, "defs": { "positive": 4, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 1, "applicable": 2 }, "additionalProperties": { "positive": 5, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 6, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/minimal_string_required.tools_argument.rendered.txt", "token_count": 442, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/minimal_string_required.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/primitive_types.tools_argument.rendered.txt", "token_count": 475, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/primitive_types.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 557, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 538, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 492, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 625, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/raw_json_schema_parameters_only.tools_argument.rendered.txt", "token_count": 441, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/function_without_parameters.tools_argument.rendered.txt", "token_count": 403, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/function_without_parameters.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 509, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 572, "style": "json schema block", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "transformed_or_collapsed_to_any", "allOf": "transformed_or_collapsed_to_any", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/empty_tools.tools_argument.rendered.txt", "token_count": 375, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/empty_tools.message_functions_json.rendered.txt", "token_count": 375, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/mistralai__Mistral-Medium-3.5-128B.analysis.json", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/mistralai__Mistral-Medium-3.5-128B.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': True, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 77 ], "snippet": "{%- set ns_c = namespace(text_parts=[], chunks=[], has_non_text=false, tool_calls=[]) %}" }, { "file": "README.md", "lines": [ 430 ], "snippet": "function_name = tool_call.function.name" }, { "file": "README.md", "lines": [ 431 ], "snippet": "function_args = tool_call.function.arguments" }, { "file": "README.md", "lines": [ 442 ], "snippet": "\"name\": tool_call.function.name," }, { "file": "chat_template.jinja", "lines": [ 90 ], "snippet": "{%- if msg['tool_calls'] is defined and msg['tool_calls'] is not none %}" }, { "file": "chat_template.jinja", "lines": [ 91 ], "snippet": "{%- set msg = {'role': msg['role'], 'content': new_content, 'tool_calls': msg['tool_calls']} %}" }, { "file": "chat_template.jinja", "lines": [ 112 ], "snippet": "{%- if msg['tool_calls'] is defined and msg['tool_calls'] is not none %}" }, { "file": "chat_template.jinja", "lines": [ 113 ], "snippet": "{%- set ns_c.tool_calls = ns_c.tool_calls + msg['tool_calls'] | list %}" }, { "file": "chat_template.jinja", "lines": [ 124 ], "snippet": "{%- if ns_c.tool_calls | length > 0 %}" } ] }, { "claim": "Rendered prompt artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'transformed_or_collapsed_to_any', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'transformed_or_collapsed_to_any', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'transformed_or_collapsed_to_any', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=442" }, { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=375" }, { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=475" }, { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=375" }, { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=557" }, { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=375" }, { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=538" }, { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=375" }, { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=492" }, { "artifact": "research/artifacts/tokenization/mistralai__Mistral-Medium-3.5-128B/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=375" } ] } ] }, "special_modes": [], "highlighted_notes": [ { "category": "instructional web-tool note", "title": "Web/fetch policy text, no builtin web tool", "detail": "The default system prompt tells the model to use dedicated tools for web search/fetch/up-to-date information and says it cannot browse without such tools. The template renders supplied tools in `[AVAILABLE_TOOLS]`; no builtin `web_search`, `web_fetch`, or browser kwarg branch was found.", "evidence": "research/artifacts/hf-snapshots/mistralai__Mistral-Medium-3.5-128B/chat_template.jinja" } ] }, { "model": "moonshotai/Kimi-K2.5", "slug": "moonshotai__Kimi-K2.5", "revision": "4d01dfe0332d63057c186e0b262165819efb6611", "support_variant": "custom_kimi_ts", "support_path": "Custom renderer path", "support_detail": "Tool support is visible through a model-specific renderer/protocol path.", "primary_style": "typescript declarations", "dialect": "typescript", "standard_supported": 0, "message_supported": 0, "custom_supported": 8, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11, "custom_kimi_ts": 11 }, "style_counts": { "typescript declarations": 8 }, "token_min": null, "token_median": null, "token_max": null, "features": { "tool_name": { "positive": 8, "applicable": 30 }, "description": { "positive": 8, "applicable": 30 }, "parameters": { "positive": 7, "applicable": 27 }, "properties": { "positive": 7, "applicable": 27 }, "required": { "positive": 7, "applicable": 27 }, "object": { "positive": 7, "applicable": 27 }, "array": { "positive": 1, "applicable": 6 }, "enum": { "positive": 2, "applicable": 9 }, "anyOf": { "positive": 1, "applicable": 6 }, "oneOf": { "positive": 0, "applicable": 3 }, "allOf": { "positive": 0, "applicable": 3 }, "ref": { "positive": 1, "applicable": 6 }, "defs": { "positive": 1, "applicable": 6 }, "default": { "positive": 1, "applicable": 6 }, "const": { "positive": 0, "applicable": 3 }, "additionalProperties": { "positive": 1, "applicable": 15 }, "minimum": { "positive": 2, "applicable": 9 }, "null": { "positive": 1, "applicable": 9 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/minimal_string_required.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/minimal_string_required.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/primitive_types.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/primitive_types.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/nested_arrays_enums.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "present_exact_key", "enum": "transformed_to_typescript_union", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "typescript declarations", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "ValueError: Invalid JSON Schema object: {'const': 'price'}" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/multiple_tools_parallel_candidate.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/refs_unions_defaults_constraints.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "present_exact_key", "additionalProperties": "omitted_in_typescript_rendering" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/raw_json_schema_parameters_only.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/function_without_parameters.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/function_without_parameters.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/non_ascii_long_descriptions.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "transformed_to_typescript_union", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted_in_typescript_rendering" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_rendered", "allOf": "not_rendered", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "typescript declarations", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_rendered", "allOf": "not_rendered", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "ValueError: Invalid JSON Schema object: {'oneOf': [{'type': 'string', 'format': 'email'}, {'type': 'string', 'pattern': '^\\\\+?[0-9-]+$'}]}" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/empty_tools.tools_argument.rendered.txt", "token_count": null, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/empty_tools.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.5/empty_tools.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/moonshotai__Kimi-K2.5/minimal_string_required.custom_kimi_ts.rendered.txt", "research/artifacts/rendered/moonshotai__Kimi-K2.5/nested_arrays_enums.custom_kimi_ts.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/moonshotai__Kimi-K2.5.analysis.json", "research/artifacts/rendered/moonshotai__Kimi-K2.5/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, has_tool_named_code_or_template, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_code", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/moonshotai__Kimi-K2.5.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': True, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': True}" }, { "file": "chat_template.jinja", "lines": [ 34 ], "snippet": "<|tool_call_begin|>{{ formatted_id }}<|tool_call_argument_begin|>{% if tool_call['function']['arguments'] is string %}{{ tool_call['function']['arguments'] }}{% else %}{{ tool_call['function']['arguments'] | tojson }}{% endif %}<|tool_call_end|>" }, { "file": "chat_template.jinja", "lines": [ 41 ], "snippet": "{%- set ns = namespace(last_non_tool_call_assistant_msg=-1) -%}" }, { "file": "chat_template.jinja", "lines": [ 31 ], "snippet": "<|tool_calls_section_begin|>" }, { "file": "chat_template.jinja", "lines": [ 32 ], "snippet": "{%- for tool_call in message['tool_calls'] -%}" }, { "file": "chat_template.jinja", "lines": [ 33 ], "snippet": "{%- set formatted_id = tool_call['id'] -%}" }, { "file": "chat_template.jinja", "lines": [ 36 ], "snippet": "<|tool_calls_section_end|>" }, { "file": "chat_template.jinja", "lines": [ 43 ], "snippet": "{%- if messages[idx]['role'] == 'assistant' and not messages[idx].get('tool_calls') -%}" }, { "file": "chat_template.jinja", "lines": [ 44 ], "snippet": "{%- set ns.last_non_tool_call_assistant_msg = idx -%}" }, { "file": "chat_template.jinja", "lines": [ 50 ], "snippet": "{%- set hist_msgs = messages[:ns.last_non_tool_call_assistant_msg+1] -%}" } ] }, { "claim": "Rendered prompt artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/minimal_string_required.custom_kimi_ts.rendered.txt", "observation": "case=minimal_string_required, method=custom_kimi_ts:custom_kimi_ts, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/primitive_types.custom_kimi_ts.rendered.txt", "observation": "case=primitive_types, method=custom_kimi_ts:custom_kimi_ts, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/nested_arrays_enums.custom_kimi_ts.rendered.txt", "observation": "case=nested_arrays_enums, method=custom_kimi_ts:custom_kimi_ts, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'transformed_to_typescript_union', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/multiple_tools_parallel_candidate.custom_kimi_ts.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=custom_kimi_ts:custom_kimi_ts, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.5/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "observation": "case=refs_unions_defaults_constraints, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" } ] } ] }, "special_modes": [], "highlighted_notes": [] }, { "model": "moonshotai/Kimi-K2.6", "slug": "moonshotai__Kimi-K2.6", "revision": "2755962d07cb42aa2d988a35bcb65cd4a9c2de82", "support_variant": "custom_kimi_ts", "support_path": "Custom renderer path", "support_detail": "Tool support is visible through a model-specific renderer/protocol path.", "primary_style": "typescript declarations", "dialect": "typescript", "standard_supported": 0, "message_supported": 0, "custom_supported": 8, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11, "custom_kimi_ts": 11 }, "style_counts": { "typescript declarations": 8 }, "token_min": null, "token_median": null, "token_max": null, "features": { "tool_name": { "positive": 8, "applicable": 30 }, "description": { "positive": 8, "applicable": 30 }, "parameters": { "positive": 7, "applicable": 27 }, "properties": { "positive": 7, "applicable": 27 }, "required": { "positive": 7, "applicable": 27 }, "object": { "positive": 7, "applicable": 27 }, "array": { "positive": 1, "applicable": 6 }, "enum": { "positive": 2, "applicable": 9 }, "anyOf": { "positive": 1, "applicable": 6 }, "oneOf": { "positive": 0, "applicable": 3 }, "allOf": { "positive": 0, "applicable": 3 }, "ref": { "positive": 1, "applicable": 6 }, "defs": { "positive": 1, "applicable": 6 }, "default": { "positive": 1, "applicable": 6 }, "const": { "positive": 0, "applicable": 3 }, "additionalProperties": { "positive": 1, "applicable": 15 }, "minimum": { "positive": 2, "applicable": 9 }, "null": { "positive": 1, "applicable": 9 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/minimal_string_required.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/minimal_string_required.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/primitive_types.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/primitive_types.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/nested_arrays_enums.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "present_exact_key", "enum": "transformed_to_typescript_union", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "typescript declarations", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "ValueError: Invalid JSON Schema object: {'const': 'price'}" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/multiple_tools_parallel_candidate.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_rendered", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_rendered", "default": "not_rendered", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/refs_unions_defaults_constraints.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "present_exact_key", "additionalProperties": "omitted_in_typescript_rendering" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/raw_json_schema_parameters_only.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/function_without_parameters.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/function_without_parameters.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/non_ascii_long_descriptions.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "transformed_to_typescript_union", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted_in_typescript_rendering" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_rendered", "allOf": "not_rendered", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; UndefinedError: 'tools_ts_str' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "typescript declarations", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_rendered", "enum": "not_rendered", "anyOf": "not_applicable", "oneOf": "not_rendered", "allOf": "not_rendered", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "ValueError: Invalid JSON Schema object: {'oneOf': [{'type': 'string', 'format': 'email'}, {'type': 'string', 'pattern': '^\\\\+?[0-9-]+$'}]}" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/empty_tools.tools_argument.rendered.txt", "token_count": null, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/empty_tools.message_functions_json.rendered.txt", "token_count": null, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType; tokenize_error: tokenizer unavailable" }, "custom_kimi_ts": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/moonshotai__Kimi-K2.6/empty_tools.custom_kimi_ts.rendered.txt", "token_count": null, "style": "typescript declarations", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/moonshotai__Kimi-K2.6/minimal_string_required.custom_kimi_ts.rendered.txt", "research/artifacts/rendered/moonshotai__Kimi-K2.6/nested_arrays_enums.custom_kimi_ts.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/moonshotai__Kimi-K2.6.analysis.json", "research/artifacts/rendered/moonshotai__Kimi-K2.6/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, has_tool_named_code_or_template, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_code", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/moonshotai__Kimi-K2.6.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': True, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': True}" }, { "file": "chat_template.jinja", "lines": [ 34 ], "snippet": "<|tool_call_begin|>{{ formatted_id }}<|tool_call_argument_begin|>{% if tool_call['function']['arguments'] is string %}{{ tool_call['function']['arguments'] }}{% else %}{{ tool_call['function']['arguments'] | tojson }}{% endif %}<|tool_call_end|>" }, { "file": "chat_template.jinja", "lines": [ 42 ], "snippet": "{%- set ns = namespace(last_non_tool_call_assistant_msg=-1) -%}" }, { "file": "chat_template.jinja", "lines": [ 31 ], "snippet": "<|tool_calls_section_begin|>" }, { "file": "chat_template.jinja", "lines": [ 32 ], "snippet": "{%- for tool_call in message['tool_calls'] -%}" }, { "file": "chat_template.jinja", "lines": [ 33 ], "snippet": "{%- set formatted_id = tool_call['id'] -%}" }, { "file": "chat_template.jinja", "lines": [ 36 ], "snippet": "<|tool_calls_section_end|>" }, { "file": "chat_template.jinja", "lines": [ 45 ], "snippet": "{%- if messages[idx]['role'] == 'assistant' and not messages[idx].get('tool_calls') -%}" }, { "file": "chat_template.jinja", "lines": [ 46 ], "snippet": "{%- set ns.last_non_tool_call_assistant_msg = idx -%}" }, { "file": "chat_template.jinja", "lines": [ 53 ], "snippet": "{%- set hist_msgs = messages[:ns.last_non_tool_call_assistant_msg+1] -%}" } ] }, { "claim": "Rendered prompt artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/minimal_string_required.custom_kimi_ts.rendered.txt", "observation": "case=minimal_string_required, method=custom_kimi_ts:custom_kimi_ts, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/primitive_types.custom_kimi_ts.rendered.txt", "observation": "case=primitive_types, method=custom_kimi_ts:custom_kimi_ts, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/nested_arrays_enums.custom_kimi_ts.rendered.txt", "observation": "case=nested_arrays_enums, method=custom_kimi_ts:custom_kimi_ts, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'transformed_to_typescript_union', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/multiple_tools_parallel_candidate.custom_kimi_ts.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=custom_kimi_ts:custom_kimi_ts, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/moonshotai__Kimi-K2.6/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "observation": "case=refs_unions_defaults_constraints, method=direct_jinja_approximation:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" } ] } ] }, "special_modes": [], "highlighted_notes": [] }, { "model": "openai/gpt-oss-120b", "slug": "openai__gpt-oss-120b", "revision": "b5c939de8f754692c1647ca79fbf85e8c1e70f8a", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "typescript declarations", "dialect": "typescript", "standard_supported": 9, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 3, "special_total": 6, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "typescript declarations": 9 }, "token_min": 79, "token_median": 79, "token_max": 182, "features": { "tool_name": { "positive": 9, "applicable": 20 }, "description": { "positive": 9, "applicable": 20 }, "parameters": { "positive": 8, "applicable": 18 }, "properties": { "positive": 8, "applicable": 18 }, "required": { "positive": 8, "applicable": 18 }, "object": { "positive": 8, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 2, "applicable": 4 }, "oneOf": { "positive": 1, "applicable": 2 }, "allOf": { "positive": 1, "applicable": 2 }, "ref": { "positive": 2, "applicable": 4 }, "defs": { "positive": 2, "applicable": 4 }, "default": { "positive": 1, "applicable": 4 }, "const": { "positive": 0, "applicable": 2 }, "additionalProperties": { "positive": 0, "applicable": 10 }, "minimum": { "positive": 0, "applicable": 6 }, "null": { "positive": 0, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/minimal_string_required.tools_argument.rendered.txt", "token_count": 135, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/minimal_string_required.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/primitive_types.tools_argument.rendered.txt", "token_count": 146, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/primitive_types.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 159, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "present_exact_key", "enum": "transformed_to_typescript_union", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 143, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "omitted_in_typescript_rendering", "additionalProperties": "omitted_in_typescript_rendering" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 153, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 159, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "not_applicable", "anyOf": "transformed_or_collapsed_to_any", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "transformed_or_collapsed_to_any", "default": "present_exact_key", "additionalProperties": "omitted_in_typescript_rendering" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": false, "evidence_grade": "not_rendered", "rendered_path": null, "token_count": null, "style": "json schema block", "features": { "tool_name": "not_rendered", "description": "not_rendered", "parameters": "not_rendered", "properties": "not_rendered", "required": "not_rendered", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_rendered" }, "error": "UndefinedError: 'dict object' has no attribute 'function'; UndefinedError: 'builtin_tools' is undefined" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/function_without_parameters.tools_argument.rendered.txt", "token_count": 120, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/function_without_parameters.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 182, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "not_applicable", "enum": "transformed_to_typescript_union", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted_in_typescript_rendering" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 141, "style": "typescript declarations", "features": { "tool_name": "present_value", "description": "transformed_to_comment", "parameters": "transformed_to_typescript_object", "properties": "transformed_to_typescript_object", "required": "transformed_to_typescript_optional_markers", "array": "present_exact_key", "enum": "transformed_to_typescript_union", "anyOf": "not_applicable", "oneOf": "transformed_or_collapsed_to_any", "allOf": "transformed_or_collapsed_to_any", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted_in_typescript_rendering" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/empty_tools.tools_argument.rendered.txt", "token_count": 79, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/empty_tools.message_functions_json.rendered.txt", "token_count": 79, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/openai__gpt-oss-120b/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/openai__gpt-oss-120b/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/openai__gpt-oss-120b/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/openai__gpt-oss-120b/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/openai__gpt-oss-120b.analysis.json", "research/artifacts/rendered/openai__gpt-oss-120b/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, builtin_special_tool_kwargs, python_tool_style, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/openai__gpt-oss-120b.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': True, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': False, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 252 ], "snippet": "{{- render_tool_namespace(\"functions\", tools) }}" }, { "file": "chat_template.jinja", "lines": [ 258 ], "snippet": "{%- set last_tool_call = namespace(name=none) %}" }, { "file": "chat_template.jinja", "lines": [ 46 ], "snippet": "{#- Handle oneOf schemas - check for complex unions and fallback to any #}" }, { "file": "chat_template.jinja", "lines": [ 108 ], "snippet": "{%- macro render_tool_namespace(namespace_name, tools) -%}" }, { "file": "chat_template.jinja", "lines": [ 112 ], "snippet": "{%- set tool = tool.function %}" }, { "file": "chat_template.jinja", "lines": [ 209 ], "snippet": "{%- set available_builtin_tools = namespace(browser=false, python=false) %}" }, { "file": "chat_template.jinja", "lines": [ 221 ], "snippet": "{{- \"\\nCalls to these tools must go to the commentary channel: 'functions'.\" }}" }, { "file": "README.md", "lines": [ 161 ], "snippet": "* Function calling with defined schemas" }, { "file": "chat_template.jinja", "lines": [ 2 ], "snippet": "In addition to the normal inputs of `messages` and `tools`, this template also accepts the" } ] }, { "claim": "Rendered prompt artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'omitted_in_typescript_rendering', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'transformed_to_typescript_union', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'omitted_in_typescript_rendering', 'allOf': 'not_applicable', 'anyOf': 'transformed_or_collapsed_to_any', 'array': 'not_applicable', 'const': 'omitted_in_typescript_rendering', 'default': 'omitted_in_typescript_rendering', 'defs': 'transformed_or_collapsed_to_any', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted_in_typescript_rendering', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'transformed_or_collapsed_to_any', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'transformed_to_typescript_type', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'transformed_to_comment', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'transformed_to_typescript_object', 'oneOf': 'not_applicable', 'parameters_key': 'transformed_to_typescript_object', 'properties_key': 'transformed_to_typescript_object', 'ref': 'not_applicable', 'required_key': 'transformed_to_typescript_optional_markers', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 21 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=135" }, { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=79" }, { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=146" }, { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=79" }, { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=159" }, { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=79" }, { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=143" }, { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=79" }, { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=153" }, { "artifact": "research/artifacts/tokenization/openai__gpt-oss-120b/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=79" } ] }, { "claim": "Builtin/special/non-JSON tool-mode artifacts were produced for 6 probe(s), separate from JSON Schema `tools=[...]` support.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Special/builtin modes are template-specific kwargs and are not equivalent to arbitrary JSON Schema function tools." ], "evidence": [ { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_python.rendered.txt", "observation": "probe=builtin_python, style=builtin_tools, method=transformers_apply_chat_template:special_tool_mode, markers=['## python', 'execute Python code', 'stateful Jupyter', 'python will respond', 'python code'], changed_from_baseline=True" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_browser.rendered.txt", "observation": "probe=builtin_browser, style=builtin_tools, method=transformers_apply_chat_template:special_tool_mode, markers=['## browser', 'namespace browser', 'type search', 'type open'], changed_from_baseline=True" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_model_identity.rendered.txt", "observation": "probe=builtin_model_identity, style=builtin_tools, method=transformers_apply_chat_template:special_tool_mode, markers=[], changed_from_baseline=True" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_reasoning_effort.rendered.txt", "observation": "probe=builtin_reasoning_effort, style=builtin_tools, method=transformers_apply_chat_template:special_tool_mode, markers=[], changed_from_baseline=True" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_medium.rendered.txt", "observation": "probe=builtin_medium, style=builtin_tools, method=transformers_apply_chat_template:special_tool_mode, markers=['medium', 'medium'], changed_from_baseline=True" }, { "artifact": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_browser_python.rendered.txt", "observation": "probe=builtin_browser_python, style=builtin_tools, method=transformers_apply_chat_template:special_tool_mode, markers=['## browser', 'namespace browser', '## python', 'execute Python code'], changed_from_baseline=True" } ] } ] }, "special_modes": [ { "probe_id": "builtin_python", "style_family": "builtin_tools", "candidate_kwarg": "builtin_tools", "category": "builtin/rendered special tool", "counts_as_special_tool": true, "description": "Rendered evidence for a non-standard tool kwarg or builtin tool mode.", "meaningful": true, "changed_from_baseline": true, "markers_found": "## python, execute Python code, stateful Jupyter, python will respond, python code", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_python.rendered.txt", "token_count": 216 }, { "probe_id": "builtin_browser", "style_family": "builtin_tools", "candidate_kwarg": "builtin_tools", "category": "builtin/rendered special tool", "counts_as_special_tool": true, "description": "Rendered evidence for a non-standard tool kwarg or builtin tool mode.", "meaningful": true, "changed_from_baseline": true, "markers_found": "## browser, namespace browser, type search, type open", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_browser.rendered.txt", "token_count": 479 }, { "probe_id": "builtin_model_identity", "style_family": "builtin_tools", "candidate_kwarg": "builtin_tools", "category": "prompt setting", "counts_as_special_tool": false, "description": "Prompt/model-setting kwarg probe; useful context, but not a callable tool.", "meaningful": true, "changed_from_baseline": true, "markers_found": "", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_model_identity.rendered.txt", "token_count": 82 }, { "probe_id": "builtin_reasoning_effort", "style_family": "builtin_tools", "candidate_kwarg": "builtin_tools", "category": "prompt setting", "counts_as_special_tool": false, "description": "Prompt/model-setting kwarg probe; useful context, but not a callable tool.", "meaningful": true, "changed_from_baseline": true, "markers_found": "", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_reasoning_effort.rendered.txt", "token_count": 82 }, { "probe_id": "builtin_medium", "style_family": "builtin_tools", "candidate_kwarg": "builtin_tools", "category": "prompt setting", "counts_as_special_tool": false, "description": "Prompt/model-setting kwarg probe; useful context, but not a callable tool.", "meaningful": true, "changed_from_baseline": true, "markers_found": "medium, medium", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_medium.rendered.txt", "token_count": 82 }, { "probe_id": "builtin_browser_python", "style_family": "builtin_tools", "candidate_kwarg": "builtin_tools", "category": "builtin/rendered special tool", "counts_as_special_tool": true, "description": "Rendered evidence for a non-standard tool kwarg or builtin tool mode.", "meaningful": true, "changed_from_baseline": true, "markers_found": "## browser, namespace browser, ## python, execute Python code", "rendered_path": "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_browser_python.rendered.txt", "token_count": 613 } ], "highlighted_notes": [] }, { "model": "zai-org/GLM-5", "slug": "zai-org__GLM-5", "revision": "4e6698ba8e85059d749020e3c4d2123719f23926", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "xml/tag syntax", "dialect": "xml", "standard_supported": 10, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "xml/tag syntax": 10 }, "token_min": 17, "token_median": 17, "token_max": 350, "features": { "tool_name": { "positive": 10, "applicable": 20 }, "description": { "positive": 10, "applicable": 20 }, "parameters": { "positive": 9, "applicable": 18 }, "properties": { "positive": 9, "applicable": 18 }, "required": { "positive": 9, "applicable": 18 }, "object": { "positive": 9, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 2, "applicable": 4 }, "oneOf": { "positive": 1, "applicable": 2 }, "allOf": { "positive": 1, "applicable": 2 }, "ref": { "positive": 2, "applicable": 4 }, "defs": { "positive": 2, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 1, "applicable": 2 }, "additionalProperties": { "positive": 5, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 3, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/minimal_string_required.tools_argument.rendered.txt", "token_count": 177, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/minimal_string_required.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/primitive_types.tools_argument.rendered.txt", "token_count": 210, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/primitive_types.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 291, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 268, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 226, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 350, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/raw_json_schema_parameters_only.tools_argument.rendered.txt", "token_count": 173, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/function_without_parameters.tools_argument.rendered.txt", "token_count": 136, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/function_without_parameters.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 247, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 300, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/empty_tools.tools_argument.rendered.txt", "token_count": 17, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5/empty_tools.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/zai-org__GLM-5/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/zai-org__GLM-5/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/zai-org__GLM-5/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/zai-org__GLM-5/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/zai-org__GLM-5.analysis.json", "research/artifacts/rendered/zai-org__GLM-5/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, openai_style_tool_mentions, custom_tool_tags_or_channels, xml_tool_style, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/zai-org__GLM-5.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': False, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 16 ], "snippet": "{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...{%- endif -%}" }, { "file": "chat_template.jinja", "lines": [ 8 ], "snippet": "You are provided with function signatures within XML tags:" }, { "file": "chat_template.jinja", "lines": [ 60 ], "snippet": "{% if m.tool_calls %}" }, { "file": "chat_template.jinja", "lines": [ 61 ], "snippet": "{% for tc in m.tool_calls %}" }, { "file": "chat_template.jinja", "lines": [ 65 ], "snippet": "{{- '' + tc.name -}}" }, { "file": "chat_template.jinja", "lines": [ 66 ], "snippet": "{% set _args = tc.arguments %}{% for k, v in _args.items() %}{{ k }}{{ v | tojson(ensure_ascii=False) if v is not string else v }}{% endfor %}{% endfor %}" }, { "file": "chat_template.jinja", "lines": [ 2 ], "snippet": "{%- if tools -%}" }, { "file": "chat_template.jinja", "lines": [ 4 ], "snippet": "# Tools" }, { "file": "chat_template.jinja", "lines": [ 6 ], "snippet": "You may call one or more functions to assist with the user query." } ] }, { "claim": "Rendered prompt artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/zai-org__GLM-5/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=177" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=17" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=210" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=17" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=291" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=17" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=268" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=17" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=226" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=17" } ] } ] }, "special_modes": [], "highlighted_notes": [] }, { "model": "zai-org/GLM-5.1", "slug": "zai-org__GLM-5.1", "revision": "442b5279043b5f43f382003e7272731f680e1393", "support_variant": "tools_argument", "support_path": "Standard tools argument", "support_detail": "`tools=[...]` renders meaningful arbitrary function schemas.", "primary_style": "xml/tag syntax", "dialect": "xml", "standard_supported": 10, "message_supported": 0, "custom_supported": 0, "case_count": 10, "special_supported": 0, "special_total": 0, "variant_counts": { "tools_argument": 11, "message_functions_json": 11 }, "style_counts": { "xml/tag syntax": 10 }, "token_min": 17, "token_median": 17, "token_max": 340, "features": { "tool_name": { "positive": 10, "applicable": 20 }, "description": { "positive": 10, "applicable": 20 }, "parameters": { "positive": 9, "applicable": 18 }, "properties": { "positive": 9, "applicable": 18 }, "required": { "positive": 9, "applicable": 18 }, "object": { "positive": 9, "applicable": 18 }, "array": { "positive": 2, "applicable": 4 }, "enum": { "positive": 3, "applicable": 6 }, "anyOf": { "positive": 2, "applicable": 4 }, "oneOf": { "positive": 1, "applicable": 2 }, "allOf": { "positive": 1, "applicable": 2 }, "ref": { "positive": 2, "applicable": 4 }, "defs": { "positive": 2, "applicable": 4 }, "default": { "positive": 2, "applicable": 4 }, "const": { "positive": 1, "applicable": 2 }, "additionalProperties": { "positive": 5, "applicable": 10 }, "minimum": { "positive": 3, "applicable": 6 }, "null": { "positive": 3, "applicable": 6 } }, "cases": { "minimal_string_required": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/minimal_string_required.tools_argument.rendered.txt", "token_count": 167, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/minimal_string_required.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "primitive_types": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/primitive_types.tools_argument.rendered.txt", "token_count": 200, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/primitive_types.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nested_arrays_enums": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/nested_arrays_enums.tools_argument.rendered.txt", "token_count": 281, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/nested_arrays_enums.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "nullable_anyof_refs": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/nullable_anyof_refs.tools_argument.rendered.txt", "token_count": 258, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/nullable_anyof_refs.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "multiple_tools_parallel_candidate": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "token_count": 206, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "refs_unions_defaults_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/refs_unions_defaults_constraints.tools_argument.rendered.txt", "token_count": 340, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "present_exact_key", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "present_exact_key", "default": "present_exact_key", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/refs_unions_defaults_constraints.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "omitted", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "omitted", "default": "omitted", "additionalProperties": "omitted" }, "error": "" } }, "raw_json_schema_parameters_only": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/raw_json_schema_parameters_only.tools_argument.rendered.txt", "token_count": 173, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/raw_json_schema_parameters_only.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "function_without_parameters": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/function_without_parameters.tools_argument.rendered.txt", "token_count": 127, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/function_without_parameters.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } }, "non_ascii_long_descriptions": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/non_ascii_long_descriptions.tools_argument.rendered.txt", "token_count": 237, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "not_applicable", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/non_ascii_long_descriptions.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "not_applicable", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "oneof_allof_constraints": { "tools_argument": { "meaningful": true, "evidence_grade": "meaningful_intended_renderer", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/oneof_allof_constraints.tools_argument.rendered.txt", "token_count": 290, "style": "xml/tag syntax", "features": { "tool_name": "present_value", "description": "present_exact_key", "parameters": "present_exact_key", "properties": "present_exact_key", "required": "present_exact_key", "array": "present_exact_key", "enum": "present_exact_key", "anyOf": "not_applicable", "oneOf": "present_exact_key", "allOf": "present_exact_key", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "present_exact_key" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "prompt_rendered_schema_missing_or_failed", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/oneof_allof_constraints.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "omitted", "description": "omitted", "parameters": "omitted", "properties": "omitted", "required": "omitted", "array": "omitted", "enum": "omitted", "anyOf": "not_applicable", "oneOf": "omitted", "allOf": "omitted", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "omitted" }, "error": "" } }, "empty_tools": { "tools_argument": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/empty_tools.tools_argument.rendered.txt", "token_count": 17, "style": "json schema block", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" }, "message_functions_json": { "meaningful": false, "evidence_grade": "empty_tools_not_applicable", "rendered_path": "research/artifacts/rendered/zai-org__GLM-5.1/empty_tools.message_functions_json.rendered.txt", "token_count": 17, "style": "message functions json", "features": { "tool_name": "not_applicable", "description": "not_applicable", "parameters": "not_applicable", "properties": "not_applicable", "required": "not_applicable", "array": "not_applicable", "enum": "not_applicable", "anyOf": "not_applicable", "oneOf": "not_applicable", "allOf": "not_applicable", "ref": "not_applicable", "default": "not_applicable", "additionalProperties": "not_applicable" }, "error": "" } } }, "representative_paths": [ "research/artifacts/rendered/zai-org__GLM-5.1/minimal_string_required.tools_argument.rendered.txt", "research/artifacts/rendered/zai-org__GLM-5.1/nested_arrays_enums.tools_argument.rendered.txt", "research/artifacts/rendered/zai-org__GLM-5.1/nullable_anyof_refs.tools_argument.rendered.txt", "research/artifacts/rendered/zai-org__GLM-5.1/oneof_allof_constraints.tools_argument.rendered.txt" ], "findings_packet": { "artifacts": [ "research/artifacts/analysis/zai-org__GLM-5.1.analysis.json", "research/artifacts/rendered/zai-org__GLM-5.1/render_summary.json" ], "claims": [ { "claim": "Static artifacts suggest: has_jinja_or_chat_template_file, openai_style_tool_mentions, typescript_style_tool_mentions, custom_tool_tags_or_channels, xml_tool_style, channel_or_special_token_style", "confidence": "medium", "evidence_class": "static_template", "limitations": [ "Automated static classification; requires human verification of cited lines." ], "evidence": [ { "artifact": "research/artifacts/analysis/zai-org__GLM-5.1.analysis.json", "observation": "source_classifications={'builtin_or_special_tool_kwargs': False, 'custom_encoding_static_only': False, 'documented_readme_support': True, 'jinja_template_tools_branch': True, 'non_json_tool_style_hints': True, 'processor_static_only': False, 'tokenizer_config_chat_template_support': False, 'tool_named_code_or_template': False}" }, { "file": "chat_template.jinja", "lines": [ 33 ], "snippet": "{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...{%- endif -%}" }, { "file": "chat_template.jinja", "lines": [ 4 ], "snippet": "{%- set ns_tool = namespace(first=true) -%}" }, { "file": "chat_template.jinja", "lines": [ 20 ], "snippet": "You are provided with function signatures within XML tags:" }, { "file": "chat_template.jinja", "lines": [ 23 ], "snippet": "{%- if 'function' in tool -%}" }, { "file": "chat_template.jinja", "lines": [ 24 ], "snippet": "{%- set tool = tool['function'] -%}" }, { "file": "chat_template.jinja", "lines": [ 83 ], "snippet": "{% if m.tool_calls %}" }, { "file": "chat_template.jinja", "lines": [ 84 ], "snippet": "{% for tc in m.tool_calls %}" }, { "file": "chat_template.jinja", "lines": [ 88 ], "snippet": "{{- '' + tc.name -}}" }, { "file": "chat_template.jinja", "lines": [ 89 ], "snippet": "{% set _args = tc.arguments %}{% for k, v in _args.items() %}{{ k }}{{ v | tojson(ensure_ascii=False) if v is not string else v }}{% endfor %}{% endfor %}" } ] }, { "claim": "Rendered prompt artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "rendered_prompt", "limitations": [ "Direct Jinja fallback, when used, is approximate and not equivalent to tokenizer apply_chat_template." ], "evidence": [ { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/minimal_string_required.tools_argument.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/minimal_string_required.message_functions_json.rendered.txt", "observation": "case=minimal_string_required, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/primitive_types.tools_argument.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'present_exact_key', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'present_exact_key', 'type_integer': 'present_exact_key', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/primitive_types.message_functions_json.rendered.txt", "observation": "case=primitive_types, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'omitted', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'omitted', 'type_integer': 'omitted', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/nested_arrays_enums.tools_argument.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'present_exact_key', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'present_exact_key', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/nested_arrays_enums.message_functions_json.rendered.txt", "observation": "case=nested_arrays_enums, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'omitted', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'omitted', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/nullable_anyof_refs.tools_argument.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'present_exact_key', 'allOf': 'not_applicable', 'anyOf': 'present_exact_key', 'array': 'not_applicable', 'const': 'present_exact_key', 'default': 'present_exact_key', 'defs': 'present_exact_key', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'present_exact_key', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'present_exact_key', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'present_exact_key', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/nullable_anyof_refs.message_functions_json.rendered.txt", "observation": "case=nullable_anyof_refs, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'omitted', 'allOf': 'not_applicable', 'anyOf': 'omitted', 'array': 'not_applicable', 'const': 'omitted', 'default': 'omitted', 'defs': 'omitted', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'omitted', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'omitted', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'omitted', 'type_string': 'omitted'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/multiple_tools_parallel_candidate.tools_argument.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:tools_argument, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'present_exact_key', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'present_exact_key', 'oneOf': 'not_applicable', 'parameters_key': 'present_exact_key', 'properties_key': 'present_exact_key', 'ref': 'not_applicable', 'required_key': 'present_exact_key', 'tool_name': 'present_value', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'present_exact_key'}" }, { "artifact": "research/artifacts/rendered/zai-org__GLM-5.1/multiple_tools_parallel_candidate.message_functions_json.rendered.txt", "observation": "case=multiple_tools_parallel_candidate, method=transformers_apply_chat_template:message_functions_json, feature_statuses={'additionalProperties': 'not_applicable', 'allOf': 'not_applicable', 'anyOf': 'not_applicable', 'array': 'not_applicable', 'const': 'not_applicable', 'default': 'not_applicable', 'defs': 'not_applicable', 'description': 'omitted', 'enum_key': 'not_applicable', 'minimum': 'not_applicable', 'null': 'not_applicable', 'object': 'omitted', 'oneOf': 'not_applicable', 'parameters_key': 'omitted', 'properties_key': 'omitted', 'ref': 'not_applicable', 'required_key': 'omitted', 'tool_name': 'omitted', 'type_boolean': 'not_applicable', 'type_integer': 'not_applicable', 'type_number': 'not_applicable', 'type_string': 'omitted'}" } ] }, { "claim": "Tokenizer artifacts were produced for 22 schema cases.", "confidence": "medium", "evidence_class": "tokenizer", "limitations": [ "Tokenization artifacts show prompt encoding only; they do not demonstrate runtime model tool-use obedience." ], "evidence": [ { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/minimal_string_required.tools_argument.tokens.json", "observation": "case=minimal_string_required, tokens=167" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/minimal_string_required.message_functions_json.tokens.json", "observation": "case=minimal_string_required, tokens=17" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/primitive_types.tools_argument.tokens.json", "observation": "case=primitive_types, tokens=200" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/primitive_types.message_functions_json.tokens.json", "observation": "case=primitive_types, tokens=17" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/nested_arrays_enums.tools_argument.tokens.json", "observation": "case=nested_arrays_enums, tokens=281" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/nested_arrays_enums.message_functions_json.tokens.json", "observation": "case=nested_arrays_enums, tokens=17" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/nullable_anyof_refs.tools_argument.tokens.json", "observation": "case=nullable_anyof_refs, tokens=258" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/nullable_anyof_refs.message_functions_json.tokens.json", "observation": "case=nullable_anyof_refs, tokens=17" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/multiple_tools_parallel_candidate.tools_argument.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=206" }, { "artifact": "research/artifacts/tokenization/zai-org__GLM-5.1/multiple_tools_parallel_candidate.message_functions_json.tokens.json", "observation": "case=multiple_tools_parallel_candidate, tokens=17" } ] } ] }, "special_modes": [], "highlighted_notes": [] } ], "cases": [ { "id": "minimal_string_required", "label": "minimal string required", "description": "Single required string argument.", "tool_names": [ "lookup_city" ], "schema": { "id": "minimal_string_required", "description": "Single required string argument.", "tools": [ { "type": "function", "function": { "name": "lookup_city", "description": "Look up information about a city.", "parameters": { "type": "object", "properties": { "city": { "type": "string", "description": "City name." } }, "required": [ "city" ] } } } ] } }, { "id": "primitive_types", "label": "primitive types", "description": "String, integer, number, and boolean fields.", "tool_names": [ "book_room" ], "schema": { "id": "primitive_types", "description": "String, integer, number, and boolean fields.", "tools": [ { "type": "function", "function": { "name": "book_room", "description": "Book a hotel room.", "parameters": { "type": "object", "properties": { "guest_name": { "type": "string" }, "nights": { "type": "integer", "minimum": 1 }, "max_price": { "type": "number" }, "breakfast": { "type": "boolean" } }, "required": [ "guest_name", "nights" ] } } } ] } }, { "id": "nested_arrays_enums", "label": "nested arrays enums", "description": "Nested objects, arrays, and string enum.", "tool_names": [ "create_ticket" ], "schema": { "id": "nested_arrays_enums", "description": "Nested objects, arrays, and string enum.", "tools": [ { "type": "function", "function": { "name": "create_ticket", "description": "Create a support ticket.", "parameters": { "type": "object", "properties": { "priority": { "type": "string", "enum": [ "low", "medium", "high" ] }, "tags": { "type": "array", "items": { "type": "string" } }, "customer": { "type": "object", "properties": { "id": { "type": "string" }, "contacts": { "type": "array", "items": { "type": "object", "properties": { "kind": { "type": "string", "enum": [ "email", "phone" ] }, "value": { "type": "string" } }, "required": [ "kind", "value" ] } } }, "required": [ "id" ] } }, "required": [ "priority", "customer" ] } } } ] } }, { "id": "nullable_anyof_refs", "label": "nullable anyof refs", "description": "Nullable/union, anyOf, $ref, defaults, and additionalProperties.", "tool_names": [ "search_inventory" ], "schema": { "id": "nullable_anyof_refs", "description": "Nullable/union, anyOf, $ref, defaults, and additionalProperties.", "tools": [ { "type": "function", "function": { "name": "search_inventory", "description": "Search inventory with advanced filters.", "parameters": { "$defs": { "priceRange": { "type": "object", "properties": { "min": { "type": "number", "default": 0 }, "max": { "type": [ "number", "null" ] } } } }, "type": "object", "properties": { "query": { "type": "string" }, "price": { "$ref": "#/$defs/priceRange" }, "sort": { "anyOf": [ { "const": "price" }, { "const": "rating" } ] }, "metadata": { "type": "object", "additionalProperties": { "type": "string" } } }, "required": [ "query" ] } } } ] } }, { "id": "multiple_tools_parallel_candidate", "label": "multiple tools parallel candidate", "description": "Multiple tools to inspect template framing and possible parallel call instructions.", "tool_names": [ "get_weather", "get_time" ], "schema": { "id": "multiple_tools_parallel_candidate", "description": "Multiple tools to inspect template framing and possible parallel call instructions.", "tools": [ { "type": "function", "function": { "name": "get_weather", "description": "Get weather for a location.", "parameters": { "type": "object", "properties": { "location": { "type": "string" } }, "required": [ "location" ] } } }, { "type": "function", "function": { "name": "get_time", "description": "Get local time for a location.", "parameters": { "type": "object", "properties": { "location": { "type": "string" } }, "required": [ "location" ] } } } ] } }, { "id": "refs_unions_defaults_constraints", "label": "refs unions defaults constraints", "description": "Advanced-but-common JSON Schema features without const/oneOf/allOf: $defs, $ref, anyOf with parseable branches, nullable type unions, defaults, constraints, and typed additionalProperties.", "tool_names": [ "plan_shipment" ], "schema": { "id": "refs_unions_defaults_constraints", "description": "Advanced-but-common JSON Schema features without const/oneOf/allOf: $defs, $ref, anyOf with parseable branches, nullable type unions, defaults, constraints, and typed additionalProperties.", "tools": [ { "type": "function", "function": { "name": "plan_shipment", "description": "Plan a shipment with reusable address and option schemas.", "parameters": { "$defs": { "address": { "type": "object", "description": "Postal address.", "properties": { "street": { "type": "string", "minLength": 1 }, "postal_code": { "type": "string", "pattern": "^[0-9A-Z -]+$" } }, "required": [ "street", "postal_code" ] } }, "type": "object", "properties": { "destination": { "$ref": "#/$defs/address" }, "weight_kg": { "type": "number", "minimum": 0.1, "default": 1.0 }, "delivery_window": { "anyOf": [ { "type": "string", "description": "Named delivery window." }, { "type": [ "number", "null" ], "description": "Unix timestamp or null." } ] }, "metadata": { "type": "object", "additionalProperties": { "type": "string" } } }, "required": [ "destination", "weight_kg" ], "additionalProperties": false } } } ] } }, { "id": "raw_json_schema_parameters_only", "label": "raw json schema parameters only", "description": "Raw JSON Schema parameter object without an OpenAI tools wrapper; useful for manual/template-adapter tests.", "tool_names": [ null ], "schema": { "id": "raw_json_schema_parameters_only", "description": "Raw JSON Schema parameter object without an OpenAI tools wrapper; useful for manual/template-adapter tests.", "tools": [ { "name": "raw_schema_probe", "description": "Probe non-wrapper compatibility.", "parameters": { "type": "object", "properties": { "q": { "type": "string", "pattern": "^[A-Z].+" } }, "required": [ "q" ], "additionalProperties": false } } ] } }, { "id": "function_without_parameters", "label": "function without parameters", "description": "Function has no parameters field.", "tool_names": [ "ping" ], "schema": { "id": "function_without_parameters", "description": "Function has no parameters field.", "tools": [ { "type": "function", "function": { "name": "ping", "description": "Return pong." } } ] } }, { "id": "non_ascii_long_descriptions", "label": "non ascii long descriptions", "description": "Non-ASCII and longer descriptions.", "tool_names": [ "translate_note" ], "schema": { "id": "non_ascii_long_descriptions", "description": "Non-ASCII and longer descriptions.", "tools": [ { "type": "function", "function": { "name": "translate_note", "description": "Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.", "parameters": { "type": "object", "properties": { "text": { "type": "string", "description": "Text to translate; may contain unicode like naïve façade résumé." }, "target_language": { "type": "string", "enum": [ "en", "fr", "ja", "es" ] } }, "required": [ "text", "target_language" ], "additionalProperties": false } } } ] } }, { "id": "oneof_allof_constraints", "label": "oneof allof constraints", "description": "oneOf/allOf, format, minItems, maxItems, numeric enum, and null enum values.", "tool_names": [ "validate_payload" ], "schema": { "id": "oneof_allof_constraints", "description": "oneOf/allOf, format, minItems, maxItems, numeric enum, and null enum values.", "tools": [ { "type": "function", "function": { "name": "validate_payload", "description": "Validate a constrained payload.", "parameters": { "type": "object", "properties": { "contact": { "oneOf": [ { "type": "string", "format": "email" }, { "type": "string", "pattern": "^\\+?[0-9-]+$" } ] }, "scores": { "type": "array", "items": { "type": "integer", "enum": [ 1, 2, 3, null ] }, "minItems": 1, "maxItems": 5 }, "profile": { "allOf": [ { "type": "object", "properties": { "name": { "type": "string" } } }, { "type": "object", "properties": { "age": { "type": "integer", "minimum": 0 } } } ] } }, "required": [ "contact" ], "additionalProperties": false } } } ] } }, { "id": "empty_tools", "label": "empty tools", "description": "No tools provided.", "tool_names": [], "schema": { "id": "empty_tools", "description": "No tools provided.", "tools": [] } } ], "snippets": { "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/empty_tools.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/empty_tools.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/function_without_parameters.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/function_without_parameters.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'ping', 'description': 'Return pong.'}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/minimal_string_required.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/minimal_string_required.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'lookup_city', 'description': 'Look up information about a city.', 'parameters': {'type': 'object', 'properties': {'city': {'type': 'string', 'description': 'City name.'}}, 'required': ['city']}}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'get_weather', 'description': 'Get weather for a location.', 'parameters': {'type': 'object', 'properties': {'location': {'type': 'string'}}, 'required': ['location']}}}\n{'type': 'function', 'function': {'name': 'get_time', 'description': 'Get local time for a location.', 'parameters': {'type': 'object', 'properties': {'location': {'type': 'string'}}, 'required': ['location']}}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nested_arrays_enums.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nested_arrays_enums.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'create_ticket', 'description': 'Create a support ticket.', 'parameters': {'type': 'object', 'properties': {'priority': {'type': 'string', 'enum': ['low', 'medium', 'high']}, 'tags': {'type': 'array', 'items': {'type': 'string'}}, 'customer': {'type': 'object', 'properties': {'id': {'type': 'string'}, 'contacts': {'type': 'array', 'items': {'type': 'object', 'properties': {'kind': {'type': 'string', 'enum': ['email', 'phone']}, 'value': {'type': 'string'}}, 'required': ['kind', 'value']}}}, 'required': ['id']}}, 'required': ['priority', 'customer']}}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/non_ascii_long_descriptions.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/non_ascii_long_descriptions.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'translate_note', 'description': 'Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.', 'parameters': {'type': 'object', 'properties': {'text': {'type': 'string', 'description': 'Text to translate; may contain unicode like naïve façade résumé.'}, 'target_language': {'type': 'string', 'enum': ['en', 'fr', 'ja', 'es']}}, 'required': ['text', 'target_language'], 'additionalProperties': False}}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nullable_anyof_refs.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/nullable_anyof_refs.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'search_inventory', 'description': 'Search inventory with advanced filters.', 'parameters': {'$defs': {'priceRange': {'type': 'object', 'properties': {'min': {'type': 'number', 'default': 0}, 'max': {'type': ['number', 'null']}}}}, 'type': 'object', 'properties': {'query': {'type': 'string'}, 'price': {'$ref': '#/$defs/priceRange'}, 'sort': {'anyOf': [{'const': 'price'}, {'const': 'rating'}]}, 'metadata': {'type': 'object', 'additionalProperties': {'type': 'string'}}}, 'required': ['query']}}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/oneof_allof_constraints.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/oneof_allof_constraints.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'validate_payload', 'description': 'Validate a constrained payload.', 'parameters': {'type': 'object', 'properties': {'contact': {'oneOf': [{'type': 'string', 'format': 'email'}, {'type': 'string', 'pattern': '^\\\\+?[0-9-]+$'}]}, 'scores': {'type': 'array', 'items': {'type': 'integer', 'enum': [1, 2, 3, None]}, 'minItems': 1, 'maxItems': 5}, 'profile': {'allOf': [{'type': 'object', 'properties': {'name': {'type': 'string'}}}, {'type': 'object', 'properties': {'age': {'type': 'integer', 'minimum': 0}}}]}}, 'required': ['contact'], 'additionalProperties': False}}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/primitive_types.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/primitive_types.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'book_room', 'description': 'Book a hotel room.', 'parameters': {'type': 'object', 'properties': {'guest_name': {'type': 'string'}, 'nights': {'type': 'integer', 'minimum': 1}, 'max_price': {'type': 'number'}, 'breakfast': {'type': 'boolean'}}, 'required': ['guest_name', 'nights']}}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/raw_json_schema_parameters_only.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'name': 'raw_schema_probe', 'description': 'Probe non-wrapper compatibility.', 'parameters': {'type': 'object', 'properties': {'q': {'type': 'string', 'pattern': '^[A-Z].+'}}, 'required': ['q'], 'additionalProperties': False}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/refs_unions_defaults_constraints.tools_argument.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'plan_shipment', 'description': 'Plan a shipment with reusable address and option schemas.', 'parameters': {'$defs': {'address': {'type': 'object', 'description': 'Postal address.', 'properties': {'street': {'type': 'string', 'minLength': 1}, 'postal_code': {'type': 'string', 'pattern': '^[0-9A-Z -]+$'}}, 'required': ['street', 'postal_code']}}, 'type': 'object', 'properties': {'destination': {'$ref': '#/$defs/address'}, 'weight_kg': {'type': 'number', 'minimum': 0.1, 'default': 1.0}, 'delivery_window': {'anyOf': [{'type': 'string', 'description': 'Named delivery window.'}, {'type': ['number', 'null'], 'description': 'Unix timestamp or null.'}]}, 'metadata': {'type': 'object', 'additionalProperties': {'type': 'string'}}}, 'required': ['destination', 'weight_kg'], 'additionalProperties': False}}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/special_tool_modes/python_tools_minimal.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nWhen you send a message containing Python code between '' and '' tags, it will be executed in a stateful Jupyter notebook environment, and you will then be given the output to continued reasoning in an agentic loop.\n\nYou can use the following tools in your python code like regular functions:\n\n{'name': 'python', 'description': 'Execute Python code.'}\n\n\nThe state persists between code executions: so variables that you define in one step are still available thereafter.\n\n<|im_end|>\n<|im_start|>user\nSolve 2+2. Use special tools if appropriate.<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/special_tool_modes/python_tools_openai_wrapped.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nWhen you send a message containing Python code between '' and '' tags, it will be executed in a stateful Jupyter notebook environment, and you will then be given the output to continued reasoning in an agentic loop.\n\nYou can use the following tools in your python code like regular functions:\n\n{'type': 'function', 'function': {'name': 'calculate', 'description': 'Evaluate a Python expression.', 'parameters': {'type': 'object', 'properties': {'expr': {'type': 'string', 'description': 'Expression to evaluate.'}}, 'required': ['expr']}}}\n\n\nThe state persists between code executions: so variables that you define in one step are still available thereafter.\n\n<|im_end|>\n<|im_start|>user\nSolve 2+2. Use special tools if appropriate.<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/HuggingFaceTB__SmolLM3-3B/special_tool_modes/xml_tools_minimal.rendered.txt": "<|im_start|>system\n## Metadata\n\nKnowledge Cutoff Date: June 2025\nToday Date: 08 May 2026\nReasoning Mode: /think\n\n## Custom Instructions\n\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face. Your role as an assistant involves thoroughly exploring questions through a systematic thinking process before providing the final precise and accurate solutions. This requires engaging in a comprehensive cycle of analysis, summarizing, exploration, reassessment, reflection, backtracking, and iteration to develop well-considered thinking process. Please structure your response into two main sections: Thought and Solution using the specified format: Thought section Solution section. In the Thought section, detail your reasoning process in steps. Each step should include detailed considerations such as analysing questions, summarizing relevant findings, brainstorming new ideas, verifying the accuracy of the current steps, refining any errors, and revisiting previous steps. In the Solution section, based on various attempts, explorations, and reflections from the Thought section, systematically present the final solution that you deem correct. The Solution section should be logical, accurate, and concise and detail necessary steps needed to reach the conclusion.\n\n### Tools\n\nYou may call one or more functions to assist with the user query.\nYou are provided with function signatures within XML tags:\n\n\n{'type': 'function', 'function': {'name': 'lookup_city', 'description': 'Look up information about a city.', 'parameters': {'type': 'object', 'properties': {'city': {'type': 'string'}}, 'required': ['city']}}}\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n\n\n<|im_end|>\n<|im_start|>user\nSolve 2+2. Use special tools if appropriate.<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/empty_tools.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/empty_tools.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/function_without_parameters.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/function_without_parameters.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"ping\", \"description\": \"Return pong.\"}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/minimal_string_required.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/minimal_string_required.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"lookup_city\", \"description\": \"Look up information about a city.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"city\": {\"type\": \"string\", \"description\": \"City name.\"}}, \"required\": [\"city\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"get_weather\", \"description\": \"Get weather for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}\n{\"name\": \"get_time\", \"description\": \"Get local time for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nested_arrays_enums.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nested_arrays_enums.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"create_ticket\", \"description\": \"Create a support ticket.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"priority\": {\"type\": \"string\", \"enum\": [\"low\", \"medium\", \"high\"]}, \"tags\": {\"type\": \"array\", \"items\": {\"type\": \"string\"}}, \"customer\": {\"type\": \"object\", \"properties\": {\"id\": {\"type\": \"string\"}, \"contacts\": {\"type\": \"array\", \"items\": {\"type\": \"object\", \"properties\": {\"kind\": {\"type\": \"string\", \"enum\": [\"email\", \"phone\"]}, \"value\": {\"type\": \"string\"}}, \"required\": [\"kind\", \"value\"]}}}, \"required\": [\"id\"]}}, \"required\": [\"priority\", \"customer\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/non_ascii_long_descriptions.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/non_ascii_long_descriptions.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"translate_note\", \"description\": \"Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"text\": {\"type\": \"string\", \"description\": \"Text to translate; may contain unicode like naïve façade résumé.\"}, \"target_language\": {\"type\": \"string\", \"enum\": [\"en\", \"fr\", \"ja\", \"es\"]}}, \"required\": [\"text\", \"target_language\"], \"additionalProperties\": false}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nullable_anyof_refs.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/nullable_anyof_refs.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"search_inventory\", \"description\": \"Search inventory with advanced filters.\", \"parameters\": {\"$defs\": {\"priceRange\": {\"type\": \"object\", \"properties\": {\"min\": {\"type\": \"number\", \"default\": 0}, \"max\": {\"type\": [\"number\", \"null\"]}}}}, \"type\": \"object\", \"properties\": {\"query\": {\"type\": \"string\"}, \"price\": {\"$ref\": \"#/$defs/priceRange\"}, \"sort\": {\"anyOf\": [{\"const\": \"price\"}, {\"const\": \"rating\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"query\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/oneof_allof_constraints.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/oneof_allof_constraints.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"validate_payload\", \"description\": \"Validate a constrained payload.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"contact\": {\"oneOf\": [{\"type\": \"string\", \"format\": \"email\"}, {\"type\": \"string\", \"pattern\": \"^\\\\+?[0-9-]+$\"}]}, \"scores\": {\"type\": \"array\", \"items\": {\"type\": \"integer\", \"enum\": [1, 2, 3, null]}, \"minItems\": 1, \"maxItems\": 5}, \"profile\": {\"allOf\": [{\"type\": \"object\", \"properties\": {\"name\": {\"type\": \"string\"}}}, {\"type\": \"object\", \"properties\": {\"age\": {\"type\": \"integer\", \"minimum\": 0}}}]}}, \"required\": [\"contact\"], \"additionalProperties\": false}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/primitive_types.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/primitive_types.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"book_room\", \"description\": \"Book a hotel room.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"guest_name\": {\"type\": \"string\"}, \"nights\": {\"type\": \"integer\", \"minimum\": 1}, \"max_price\": {\"type\": \"number\"}, \"breakfast\": {\"type\": \"boolean\"}}, \"required\": [\"guest_name\", \"nights\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.5/refs_unions_defaults_constraints.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"plan_shipment\", \"description\": \"Plan a shipment with reusable address and option schemas.\", \"parameters\": {\"$defs\": {\"address\": {\"type\": \"object\", \"description\": \"Postal address.\", \"properties\": {\"street\": {\"type\": \"string\", \"minLength\": 1}, \"postal_code\": {\"type\": \"string\", \"pattern\": \"^[0-9A-Z -]+$\"}}, \"required\": [\"street\", \"postal_code\"]}}, \"type\": \"object\", \"properties\": {\"destination\": {\"$ref\": \"#/$defs/address\"}, \"weight_kg\": {\"type\": \"number\", \"minimum\": 0.1, \"default\": 1.0}, \"delivery_window\": {\"anyOf\": [{\"type\": \"string\", \"description\": \"Named delivery window.\"}, {\"type\": [\"number\", \"null\"], \"description\": \"Unix timestamp or null.\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"destination\", \"weight_kg\"], \"additionalProperties\": false}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/empty_tools.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/empty_tools.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/function_without_parameters.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/function_without_parameters.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"ping\", \"description\": \"Return pong.\"}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/minimal_string_required.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/minimal_string_required.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"lookup_city\", \"description\": \"Look up information about a city.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"city\": {\"type\": \"string\", \"description\": \"City name.\"}}, \"required\": [\"city\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"get_weather\", \"description\": \"Get weather for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}\n{\"name\": \"get_time\", \"description\": \"Get local time for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nested_arrays_enums.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nested_arrays_enums.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"create_ticket\", \"description\": \"Create a support ticket.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"priority\": {\"type\": \"string\", \"enum\": [\"low\", \"medium\", \"high\"]}, \"tags\": {\"type\": \"array\", \"items\": {\"type\": \"string\"}}, \"customer\": {\"type\": \"object\", \"properties\": {\"id\": {\"type\": \"string\"}, \"contacts\": {\"type\": \"array\", \"items\": {\"type\": \"object\", \"properties\": {\"kind\": {\"type\": \"string\", \"enum\": [\"email\", \"phone\"]}, \"value\": {\"type\": \"string\"}}, \"required\": [\"kind\", \"value\"]}}}, \"required\": [\"id\"]}}, \"required\": [\"priority\", \"customer\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/non_ascii_long_descriptions.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/non_ascii_long_descriptions.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"translate_note\", \"description\": \"Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"text\": {\"type\": \"string\", \"description\": \"Text to translate; may contain unicode like naïve façade résumé.\"}, \"target_language\": {\"type\": \"string\", \"enum\": [\"en\", \"fr\", \"ja\", \"es\"]}}, \"required\": [\"text\", \"target_language\"], \"additionalProperties\": false}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nullable_anyof_refs.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/nullable_anyof_refs.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"search_inventory\", \"description\": \"Search inventory with advanced filters.\", \"parameters\": {\"$defs\": {\"priceRange\": {\"type\": \"object\", \"properties\": {\"min\": {\"type\": \"number\", \"default\": 0}, \"max\": {\"type\": [\"number\", \"null\"]}}}}, \"type\": \"object\", \"properties\": {\"query\": {\"type\": \"string\"}, \"price\": {\"$ref\": \"#/$defs/priceRange\"}, \"sort\": {\"anyOf\": [{\"const\": \"price\"}, {\"const\": \"rating\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"query\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/oneof_allof_constraints.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/oneof_allof_constraints.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"validate_payload\", \"description\": \"Validate a constrained payload.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"contact\": {\"oneOf\": [{\"type\": \"string\", \"format\": \"email\"}, {\"type\": \"string\", \"pattern\": \"^\\\\+?[0-9-]+$\"}]}, \"scores\": {\"type\": \"array\", \"items\": {\"type\": \"integer\", \"enum\": [1, 2, 3, null]}, \"minItems\": 1, \"maxItems\": 5}, \"profile\": {\"allOf\": [{\"type\": \"object\", \"properties\": {\"name\": {\"type\": \"string\"}}}, {\"type\": \"object\", \"properties\": {\"age\": {\"type\": \"integer\", \"minimum\": 0}}}]}}, \"required\": [\"contact\"], \"additionalProperties\": false}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/primitive_types.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/primitive_types.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"book_room\", \"description\": \"Book a hotel room.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"guest_name\": {\"type\": \"string\"}, \"nights\": {\"type\": \"integer\", \"minimum\": 1}, \"max_price\": {\"type\": \"number\"}, \"breakfast\": {\"type\": \"boolean\"}}, \"required\": [\"guest_name\", \"nights\"]}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/MiniMaxAI__MiniMax-M2.7/refs_unions_defaults_constraints.tools_argument.rendered.txt": "]~!b[]~b]system\nYou are a helpful assistant. Your name is MiniMax-M2.7 and is built by MiniMax.\n\n# Tools\nYou may call one or more tools to assist with the user query.\nHere are the tools available in JSONSchema format:\n\n\n{\"name\": \"plan_shipment\", \"description\": \"Plan a shipment with reusable address and option schemas.\", \"parameters\": {\"$defs\": {\"address\": {\"type\": \"object\", \"description\": \"Postal address.\", \"properties\": {\"street\": {\"type\": \"string\", \"minLength\": 1}, \"postal_code\": {\"type\": \"string\", \"pattern\": \"^[0-9A-Z -]+$\"}}, \"required\": [\"street\", \"postal_code\"]}}, \"type\": \"object\", \"properties\": {\"destination\": {\"$ref\": \"#/$defs/address\"}, \"weight_kg\": {\"type\": \"number\", \"minimum\": 0.1, \"default\": 1.0}, \"delivery_window\": {\"anyOf\": [{\"type\": \"string\", \"description\": \"Named delivery window.\"}, {\"type\": [\"number\", \"null\"], \"description\": \"Unix timestamp or null.\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"destination\", \"weight_kg\"], \"additionalProperties\": false}}\n\n\nWhen making tool calls, use XML format to invoke tools and pass parameters:\n\n\n\nparam-value-1\nparam-value-2\n...\n\n[e~[\n]~b]user\nUse the available tools if needed: what is the answer?[e~[\n]~b]ai\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/empty_tools.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/empty_tools.tools_argument.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/function_without_parameters.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/function_without_parameters.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"ping\", \"description\": \"Return pong.\"}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/minimal_string_required.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/minimal_string_required.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"lookup_city\", \"description\": \"Look up information about a city.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"city\": {\"type\": \"string\", \"description\": \"City name.\"}}, \"required\": [\"city\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"get_weather\", \"description\": \"Get weather for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}}\n{\"type\": \"function\", \"function\": {\"name\": \"get_time\", \"description\": \"Get local time for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nested_arrays_enums.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nested_arrays_enums.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"create_ticket\", \"description\": \"Create a support ticket.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"priority\": {\"type\": \"string\", \"enum\": [\"low\", \"medium\", \"high\"]}, \"tags\": {\"type\": \"array\", \"items\": {\"type\": \"string\"}}, \"customer\": {\"type\": \"object\", \"properties\": {\"id\": {\"type\": \"string\"}, \"contacts\": {\"type\": \"array\", \"items\": {\"type\": \"object\", \"properties\": {\"kind\": {\"type\": \"string\", \"enum\": [\"email\", \"phone\"]}, \"value\": {\"type\": \"string\"}}, \"required\": [\"kind\", \"value\"]}}}, \"required\": [\"id\"]}}, \"required\": [\"priority\", \"customer\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/non_ascii_long_descriptions.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/non_ascii_long_descriptions.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"translate_note\", \"description\": \"Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"text\": {\"type\": \"string\", \"description\": \"Text to translate; may contain unicode like naïve façade résumé.\"}, \"target_language\": {\"type\": \"string\", \"enum\": [\"en\", \"fr\", \"ja\", \"es\"]}}, \"required\": [\"text\", \"target_language\"], \"additionalProperties\": false}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nullable_anyof_refs.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/nullable_anyof_refs.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"search_inventory\", \"description\": \"Search inventory with advanced filters.\", \"parameters\": {\"$defs\": {\"priceRange\": {\"type\": \"object\", \"properties\": {\"min\": {\"type\": \"number\", \"default\": 0}, \"max\": {\"type\": [\"number\", \"null\"]}}}}, \"type\": \"object\", \"properties\": {\"query\": {\"type\": \"string\"}, \"price\": {\"$ref\": \"#/$defs/priceRange\"}, \"sort\": {\"anyOf\": [{\"const\": \"price\"}, {\"const\": \"rating\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"query\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/oneof_allof_constraints.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/oneof_allof_constraints.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"validate_payload\", \"description\": \"Validate a constrained payload.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"contact\": {\"oneOf\": [{\"type\": \"string\", \"format\": \"email\"}, {\"type\": \"string\", \"pattern\": \"^\\\\+?[0-9-]+$\"}]}, \"scores\": {\"type\": \"array\", \"items\": {\"type\": \"integer\", \"enum\": [1, 2, 3, null]}, \"minItems\": 1, \"maxItems\": 5}, \"profile\": {\"allOf\": [{\"type\": \"object\", \"properties\": {\"name\": {\"type\": \"string\"}}}, {\"type\": \"object\", \"properties\": {\"age\": {\"type\": \"integer\", \"minimum\": 0}}}]}}, \"required\": [\"contact\"], \"additionalProperties\": false}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/primitive_types.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/primitive_types.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"book_room\", \"description\": \"Book a hotel room.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"guest_name\": {\"type\": \"string\"}, \"nights\": {\"type\": \"integer\", \"minimum\": 1}, \"max_price\": {\"type\": \"number\"}, \"breakfast\": {\"type\": \"boolean\"}}, \"required\": [\"guest_name\", \"nights\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/raw_json_schema_parameters_only.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"name\": \"raw_schema_probe\", \"description\": \"Probe non-wrapper compatibility.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"q\": {\"type\": \"string\", \"pattern\": \"^[A-Z].+\"}}, \"required\": [\"q\"], \"additionalProperties\": false}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.5-397B-A17B/refs_unions_defaults_constraints.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"plan_shipment\", \"description\": \"Plan a shipment with reusable address and option schemas.\", \"parameters\": {\"$defs\": {\"address\": {\"type\": \"object\", \"description\": \"Postal address.\", \"properties\": {\"street\": {\"type\": \"string\", \"minLength\": 1}, \"postal_code\": {\"type\": \"string\", \"pattern\": \"^[0-9A-Z -]+$\"}}, \"required\": [\"street\", \"postal_code\"]}}, \"type\": \"object\", \"properties\": {\"destination\": {\"$ref\": \"#/$defs/address\"}, \"weight_kg\": {\"type\": \"number\", \"minimum\": 0.1, \"default\": 1.0}, \"delivery_window\": {\"anyOf\": [{\"type\": \"string\", \"description\": \"Named delivery window.\"}, {\"type\": [\"number\", \"null\"], \"description\": \"Unix timestamp or null.\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"destination\", \"weight_kg\"], \"additionalProperties\": false}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/empty_tools.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/empty_tools.tools_argument.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/function_without_parameters.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/function_without_parameters.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"ping\", \"description\": \"Return pong.\"}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/minimal_string_required.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/minimal_string_required.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"lookup_city\", \"description\": \"Look up information about a city.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"city\": {\"type\": \"string\", \"description\": \"City name.\"}}, \"required\": [\"city\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"get_weather\", \"description\": \"Get weather for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}}\n{\"type\": \"function\", \"function\": {\"name\": \"get_time\", \"description\": \"Get local time for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nested_arrays_enums.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nested_arrays_enums.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"create_ticket\", \"description\": \"Create a support ticket.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"priority\": {\"type\": \"string\", \"enum\": [\"low\", \"medium\", \"high\"]}, \"tags\": {\"type\": \"array\", \"items\": {\"type\": \"string\"}}, \"customer\": {\"type\": \"object\", \"properties\": {\"id\": {\"type\": \"string\"}, \"contacts\": {\"type\": \"array\", \"items\": {\"type\": \"object\", \"properties\": {\"kind\": {\"type\": \"string\", \"enum\": [\"email\", \"phone\"]}, \"value\": {\"type\": \"string\"}}, \"required\": [\"kind\", \"value\"]}}}, \"required\": [\"id\"]}}, \"required\": [\"priority\", \"customer\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/non_ascii_long_descriptions.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/non_ascii_long_descriptions.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"translate_note\", \"description\": \"Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"text\": {\"type\": \"string\", \"description\": \"Text to translate; may contain unicode like naïve façade résumé.\"}, \"target_language\": {\"type\": \"string\", \"enum\": [\"en\", \"fr\", \"ja\", \"es\"]}}, \"required\": [\"text\", \"target_language\"], \"additionalProperties\": false}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nullable_anyof_refs.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/nullable_anyof_refs.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"search_inventory\", \"description\": \"Search inventory with advanced filters.\", \"parameters\": {\"$defs\": {\"priceRange\": {\"type\": \"object\", \"properties\": {\"min\": {\"type\": \"number\", \"default\": 0}, \"max\": {\"type\": [\"number\", \"null\"]}}}}, \"type\": \"object\", \"properties\": {\"query\": {\"type\": \"string\"}, \"price\": {\"$ref\": \"#/$defs/priceRange\"}, \"sort\": {\"anyOf\": [{\"const\": \"price\"}, {\"const\": \"rating\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"query\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/oneof_allof_constraints.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/oneof_allof_constraints.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"validate_payload\", \"description\": \"Validate a constrained payload.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"contact\": {\"oneOf\": [{\"type\": \"string\", \"format\": \"email\"}, {\"type\": \"string\", \"pattern\": \"^\\\\+?[0-9-]+$\"}]}, \"scores\": {\"type\": \"array\", \"items\": {\"type\": \"integer\", \"enum\": [1, 2, 3, null]}, \"minItems\": 1, \"maxItems\": 5}, \"profile\": {\"allOf\": [{\"type\": \"object\", \"properties\": {\"name\": {\"type\": \"string\"}}}, {\"type\": \"object\", \"properties\": {\"age\": {\"type\": \"integer\", \"minimum\": 0}}}]}}, \"required\": [\"contact\"], \"additionalProperties\": false}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/primitive_types.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/primitive_types.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"book_room\", \"description\": \"Book a hotel room.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"guest_name\": {\"type\": \"string\"}, \"nights\": {\"type\": \"integer\", \"minimum\": 1}, \"max_price\": {\"type\": \"number\"}, \"breakfast\": {\"type\": \"boolean\"}}, \"required\": [\"guest_name\", \"nights\"]}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/raw_json_schema_parameters_only.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"name\": \"raw_schema_probe\", \"description\": \"Probe non-wrapper compatibility.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"q\": {\"type\": \"string\", \"pattern\": \"^[A-Z].+\"}}, \"required\": [\"q\"], \"additionalProperties\": false}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/Qwen__Qwen3.6-35B-A3B/refs_unions_defaults_constraints.tools_argument.rendered.txt": "<|im_start|>system\n# Tools\n\nYou have access to the following functions:\n\n\n{\"type\": \"function\", \"function\": {\"name\": \"plan_shipment\", \"description\": \"Plan a shipment with reusable address and option schemas.\", \"parameters\": {\"$defs\": {\"address\": {\"type\": \"object\", \"description\": \"Postal address.\", \"properties\": {\"street\": {\"type\": \"string\", \"minLength\": 1}, \"postal_code\": {\"type\": \"string\", \"pattern\": \"^[0-9A-Z -]+$\"}}, \"required\": [\"street\", \"postal_code\"]}}, \"type\": \"object\", \"properties\": {\"destination\": {\"$ref\": \"#/$defs/address\"}, \"weight_kg\": {\"type\": \"number\", \"minimum\": 0.1, \"default\": 1.0}, \"delivery_window\": {\"anyOf\": [{\"type\": \"string\", \"description\": \"Named delivery window.\"}, {\"type\": [\"number\", \"null\"], \"description\": \"Unix timestamp or null.\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"destination\", \"weight_kg\"], \"additionalProperties\": false}}}\n\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/empty_tools.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/empty_tools.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/function_without_parameters.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/function_without_parameters.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\nping\nReturn pong.\n\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/minimal_string_required.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/minimal_string_required.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\nlookup_city\nLook up information about a city.\n\n\ncity\nstring\nCity name.\n\n[\"city\"]\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\nget_weather\nGet weather for a location.\n\n\nlocation\nstring\n\n[\"location\"]\n\n\n\nget_time\nGet local time for a location.\n\n\nlocation\nstring\n\n[\"location\"]\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nested_arrays_enums.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nested_arrays_enums.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\ncreate_ticket\nCreate a support ticket.\n\n\npriority\nstring\n[\"low\", \"medium\", \"high\"]\n\n\ntags\narray\n{\"type\": \"string\"}\n\n\ncustomer\nobject\n{\"id\": {\"type\": \"string\"}, \"contacts\": {\"type\": \"array\", \"items\": {\"type\": \"object\", \"properties\": {\"kind\": {\"type\": \"string\", \"enum\": [\"email\", \"phone\"]}, \"value\": {\"type\": \"string\"}}, \"required\": [\"kind\", \"value\"]}}}\n[\"id\"]\n\n[\"priority\", \"customer\"]\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/non_ascii_long_descriptions.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/non_ascii_long_descriptions.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\ntranslate_note\nTranslate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\n\n\ntext\nstring\nText to translate; may contain unicode like naïve façade résumé.\n\n\ntarget_language\nstring\n[\"en\", \"fr\", \"ja\", \"es\"]\n\n[\"text\", \"target_language\"]\nFalse\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nullable_anyof_refs.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/nullable_anyof_refs.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\nsearch_inventory\nSearch inventory with advanced filters.\n\n\nquery\nstring\n\n\nprice\n<$ref>#/$defs/priceRange\n\n\nsort\n[{\"const\": \"price\"}, {\"const\": \"rating\"}]\n\n\nmetadata\nobject\n{\"type\": \"string\"}\n\n<$defs>{\"priceRange\": {\"type\": \"object\", \"properties\": {\"min\": {\"type\": \"number\", \"default\": 0}, \"max\": {\"type\": [\"number\", \"null\"]}}}}\n[\"query\"]\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/oneof_allof_constraints.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/oneof_allof_constraints.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\nvalidate_payload\nValidate a constrained payload.\n\n\ncontact\n[{\"type\": \"string\", \"format\": \"email\"}, {\"type\": \"string\", \"pattern\": \"^\\\\+?[0-9-]+$\"}]\n\n\nscores\narray\n{\"type\": \"integer\", \"enum\": [1, 2, 3, null]}\n1\n5\n\n\nprofile\n[{\"type\": \"object\", \"properties\": {\"name\": {\"type\": \"string\"}}}, {\"type\": \"object\", \"properties\": {\"age\": {\"type\": \"integer\", \"minimum\": 0}}}]\n\n[\"contact\"]\nFalse\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/primitive_types.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/primitive_types.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\nbook_room\nBook a hotel room.\n\n\nguest_name\nstring\n\n\nnights\ninteger\n1\n\n\nmax_price\nnumber\n\n\nbreakfast\nboolean\n\n[\"guest_name\", \"nights\"]\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/raw_json_schema_parameters_only.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\nraw_schema_probe\nProbe non-wrapper compatibility.\n\n\nq\nstring\n^[A-Z].+\n\n[\"q\"]\nFalse\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/refs_unions_defaults_constraints.tools_argument.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou have access to the following functions:\n\n\n\nplan_shipment\nPlan a shipment with reusable address and option schemas.\n\n\ndestination\n<$ref>#/$defs/address\n\n\nweight_kg\nnumber\n0.1\n1.0\n\n\ndelivery_window\n[{\"type\": \"string\", \"description\": \"Named delivery window.\"}, {\"type\": [\"number\", \"null\"], \"description\": \"Unix timestamp or null.\"}]\n\n\nmetadata\nobject\n{\"type\": \"string\"}\n\n<$defs>{\"address\": {\"type\": \"object\", \"description\": \"Postal address.\", \"properties\": {\"street\": {\"type\": \"string\", \"minLength\": 1}, \"postal_code\": {\"type\": \"string\", \"pattern\": \"^[0-9A-Z -]+$\"}}, \"required\": [\"street\", \"postal_code\"]}}\n[\"destination\", \"weight_kg\"]\nFalse\n\n\n\n\nFor each function call, output the function name and arguments in the following format:\n\n\nvalue_1\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- DO NOT use function calls inside tags.\n- The value enclosed between parameter tags is preserved exactly as-is, including newlines and spaces.\n<|im_end|><|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|><|im_start|>assistant\n", "research/artifacts/rendered/XiaomiMiMo__MiMo-V2.5-Pro/special_tool_modes/ad_hoc_mimo_tool_messages.rendered.txt": "<|im_start|>system\nYou are MiMo, a warm, curious, and thoughtful AI assistant.\n\n## Identity\n\n- Your name is **MiMo-v2.5-pro**.\n- You were developed by the **Xiaomi MiMo Team**.\n- You have **1T parameters** and a **1M-token context window**.\n- When asked about your creator, identify the **Xiaomi MiMo Team**.\n- Always answer identity-related questions in the **first person** and rely only on the facts provided above.\n\n# Response Style\n\nBe warm, conversational, and respectful. Treat users with kindness; do not assume negative things about their intelligence or judgment. You may disagree, but do so constructively.\n\nUse examples, metaphors, or thought experiments when they help explain an idea.\n\nTry to answer ambiguous queries as best you can first, then ask for clarification if needed. When you do ask, limit yourself to one question at a time.\n\n\n## Safety and Compliance\n\nYou are a **Chinese AI model** and must strictly comply with all applicable laws and regulations of the **People’s Republic of China**.\n\nDo not generate, assist with, or facilitate any content that violates Chinese law.<|im_end|><|im_start|>user\nCall lookup_city for Paris, then summarize the tool result.<|im_end|><|im_start|>assistant\nNeed city lookup.\n\nParis\n\n<|im_end|><|im_start|>tool\n\n{\"city\":\"Paris\",\"country\":\"France\"}\n\n<|im_end|><|im_start|>assistant\nParis is in France.<|im_end|>", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/empty_tools.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/empty_tools.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/function_without_parameters.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"function\": {\n \"description\": \"Return pong.\",\n \"name\": \"ping\"\n },\n \"type\": \"function\"\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/function_without_parameters.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/minimal_string_required.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"function\": {\n \"description\": \"Look up information about a city.\",\n \"name\": \"lookup_city\",\n \"parameters\": {\n \"properties\": {\n \"city\": {\n \"description\": \"City name.\",\n \"type\": \"string\"\n }\n },\n \"required\": [\n \"city\"\n ],\n \"type\": \"object\"\n }\n },\n \"type\": \"function\"\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/minimal_string_required.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"function\": {\n \"description\": \"Get weather for a location.\",\n \"name\": \"get_weather\",\n \"parameters\": {\n \"properties\": {\n \"location\": {\n \"type\": \"string\"\n }\n },\n \"required\": [\n \"location\"\n ],\n \"type\": \"object\"\n }\n },\n \"type\": \"function\"\n },\n {\n \"function\": {\n \"description\": \"Get local time for a location.\",\n \"name\": \"get_time\",\n \"parameters\": {\n \"properties\": {\n \"location\": {\n \"type\": \"string\"\n }\n },\n \"required\": [\n \"location\"\n ],\n \"type\": \"object\"\n }\n },\n \"type\": \"function\"\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nested_arrays_enums.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"function\": {\n \"description\": \"Create a support ticket.\",\n \"name\": \"create_ticket\",\n \"parameters\": {\n \"properties\": {\n \"customer\": {\n \"properties\": {\n \"contacts\": {\n \"items\": {\n \"properties\": {\n \"kind\": {\n \"enum\": [\n \"email\",\n \"phone\"\n ],\n \"type\": \"string\"\n },\n \"value\": {\n \"type\": \"string\"\n }\n },\n \"required\": [\n \"kind\",\n \"value\"\n ],\n \"type\": \"object\"\n },\n \"type\": \"array\"\n },\n \"id\": {\n \"type\": \"string\"\n }\n },\n \"required\": [\n \"id\"\n ],\n \"type\": \"object\"\n },\n \"priority\": {\n \"enum\": [\n \"low\",\n \"medium\",\n \"high\"\n ],\n \"type\": \"string\"\n },\n \"tags\": {\n \"items\": {\n \"type\": \"string\"\n },\n \"type\": \"array\"\n }\n },\n \"required\": [\n \"priority\",\n \"customer\"\n ],\n \"type\": \"object\"\n }\n },\n \"type\": \"function\"\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nested_arrays_enums.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/non_ascii_long_descriptions.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"function\": {\n \"description\": \"Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\",\n \"name\": \"translate_note\",\n \"parameters\": {\n \"additionalProperties\": false,\n \"properties\": {\n \"target_language\": {\n \"enum\": [\n \"en\",\n \"fr\",\n \"ja\",\n \"es\"\n ],\n \"type\": \"string\"\n },\n \"text\": {\n \"description\": \"Text to translate; may contain unicode like naïve façade résumé.\",\n \"type\": \"string\"\n }\n },\n \"required\": [\n \"text\",\n \"target_language\"\n ],\n \"type\": \"object\"\n }\n },\n \"type\": \"function\"\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/non_ascii_long_descriptions.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nullable_anyof_refs.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"function\": {\n \"description\": \"Search inventory with advanced filters.\",\n \"name\": \"search_inventory\",\n \"parameters\": {\n \"$defs\": {\n \"priceRange\": {\n \"properties\": {\n \"max\": {\n \"type\": [\n \"number\",\n \"null\"\n ]\n },\n \"min\": {\n \"default\": 0,\n \"type\": \"number\"\n }\n },\n \"type\": \"object\"\n }\n },\n \"properties\": {\n \"metadata\": {\n \"additionalProperties\": {\n \"type\": \"string\"\n },\n \"type\": \"object\"\n },\n \"price\": {\n \"$ref\": \"#/$defs/priceRange\"\n },\n \"query\": {\n \"type\": \"string\"\n },\n \"sort\": {\n \"anyOf\": [\n {\n \"const\": \"price\"\n },\n {\n \"const\": \"rating\"\n }\n ]\n }\n },\n \"required\": [\n \"query\"\n ],\n \"type\": \"object\"\n }\n },\n \"type\": \"function\"\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/nullable_anyof_refs.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/oneof_allof_constraints.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"function\": {\n \"description\": \"Validate a constrained payload.\",\n \"name\": \"validate_payload\",\n \"parameters\": {\n \"additionalProperties\": false,\n \"properties\": {\n \"contact\": {\n \"oneOf\": [\n {\n \"format\": \"email\",\n \"type\": \"string\"\n },\n {\n \"pattern\": \"^\\\\+?[0-9-]+$\",\n \"type\": \"string\"\n }\n ]\n },\n \"profile\": {\n \"allOf\": [\n {\n \"properties\": {\n \"name\": {\n \"type\": \"string\"\n }\n },\n \"type\": \"object\"\n },\n {\n \"properties\": {\n \"age\": {\n \"minimum\": 0,\n \"type\": \"integer\"\n }\n },\n \"type\": \"object\"\n }\n ]\n },\n \"scores\": {\n \"items\": {\n \"enum\": [\n 1,\n 2,\n 3,\n null\n ],\n \"type\": \"integer\"\n },\n \"maxItems\": 5,\n \"minItems\": 1,\n \"type\": \"array\"\n }\n },\n \"required\": [\n \"contact\"\n ],\n \"type\": \"object\"\n }\n },\n \"type\": \"function\"\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/oneof_allof_constraints.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/primitive_types.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"function\": {\n \"description\": \"Book a hotel room.\",\n \"name\": \"book_room\",\n \"parameters\": {\n \"properties\": {\n \"breakfast\": {\n \"type\": \"boolean\"\n },\n \"guest_name\": {\n \"type\": \"string\"\n },\n \"max_price\": {\n \"type\": \"number\"\n },\n \"nights\": {\n \"minimum\": 1,\n \"type\": \"integer\"\n }\n },\n \"required\": [\n \"guest_name\",\n \"nights\"\n ],\n \"type\": \"object\"\n }\n },\n \"type\": \"function\"\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/primitive_types.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"description\": \"Probe non-wrapper compatibility.\",\n \"name\": \"raw_schema_probe\",\n \"parameters\": {\n \"additionalProperties\": false,\n \"properties\": {\n \"q\": {\n \"pattern\": \"^[A-Z].+\",\n \"type\": \"string\"\n }\n },\n \"required\": [\n \"q\"\n ],\n \"type\": \"object\"\n }\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/raw_json_schema_parameters_only.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?\n[\n {\n \"function\": {\n \"description\": \"Plan a shipment with reusable address and option schemas.\",\n \"name\": \"plan_shipment\",\n \"parameters\": {\n \"$defs\": {\n \"address\": {\n \"description\": \"Postal address.\",\n \"properties\": {\n \"postal_code\": {\n \"pattern\": \"^[0-9A-Z -]+$\",\n \"type\": \"string\"\n },\n \"street\": {\n \"minLength\": 1,\n \"type\": \"string\"\n }\n },\n \"required\": [\n \"street\",\n \"postal_code\"\n ],\n \"type\": \"object\"\n }\n },\n \"additionalProperties\": false,\n \"properties\": {\n \"delivery_window\": {\n \"anyOf\": [\n {\n \"description\": \"Named delivery window.\",\n \"type\": \"string\"\n },\n {\n \"description\": \"Unix timestamp or null.\",\n \"type\": [\n \"number\",\n \"null\"\n ]\n }\n ]\n },\n \"destination\": {\n \"$ref\": \"#/$defs/address\"\n },\n \"metadata\": {\n \"additionalProperties\": {\n \"type\": \"string\"\n },\n \"type\": \"object\"\n },\n \"weight_kg\": {\n \"default\": 1.0,\n \"minimum\": 0.1,\n \"type\": \"number\"\n }\n },\n \"required\": [\n \"destination\",\n \"weight_kg\"\n ],\n \"type\": \"object\"\n }\n },\n \"type\": \"function\"\n }\n]<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/allenai__Olmo-3.1-32B-Think/refs_unions_defaults_constraints.tools_argument.rendered.txt": "<|im_start|>system\nYou are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.<|im_end|>\n<|im_start|>user\nUse the available tools if needed: what is the answer?<|im_end|>\n<|im_start|>assistant\n", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/empty_tools.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/function_without_parameters.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.\n\n## Tools\n\nYou have access to a set of tools to help answer the user's question. You can invoke tools by writing a \"<｜DSML｜tool_calls>\" block like the following:\n\n<｜DSML｜tool_calls>\n<｜DSML｜invoke name=\"$TOOL_NAME\">\n<｜DSML｜parameter name=\"$PARAMETER_NAME\" string=\"true|false\">$PARAMETER_VALUE\n...\n\n<｜DSML｜invoke name=\"$TOOL_NAME2\">\n...\n\n\n\nString parameters should be specified as is and set `string=\"true\"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string=\"false\"`.\n\nIf thinking_mode is enabled (triggered by ), you MUST output your complete reasoning inside ... BEFORE any tool calls or final response.\n\nOtherwise, output directly after with tool calls or final response.\n\n### Available Tool Schemas\n\n{\"name\": \"ping\", \"description\": \"Return pong.\"}\n\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/minimal_string_required.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.\n\n## Tools\n\nYou have access to a set of tools to help answer the user's question. You can invoke tools by writing a \"<｜DSML｜tool_calls>\" block like the following:\n\n<｜DSML｜tool_calls>\n<｜DSML｜invoke name=\"$TOOL_NAME\">\n<｜DSML｜parameter name=\"$PARAMETER_NAME\" string=\"true|false\">$PARAMETER_VALUE\n...\n\n<｜DSML｜invoke name=\"$TOOL_NAME2\">\n...\n\n\n\nString parameters should be specified as is and set `string=\"true\"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string=\"false\"`.\n\nIf thinking_mode is enabled (triggered by ), you MUST output your complete reasoning inside ... BEFORE any tool calls or final response.\n\nOtherwise, output directly after with tool calls or final response.\n\n### Available Tool Schemas\n\n{\"name\": \"lookup_city\", \"description\": \"Look up information about a city.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"city\": {\"type\": \"string\", \"description\": \"City name.\"}}, \"required\": [\"city\"]}}\n\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/multiple_tools_parallel_candidate.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.\n\n## Tools\n\nYou have access to a set of tools to help answer the user's question. You can invoke tools by writing a \"<｜DSML｜tool_calls>\" block like the following:\n\n<｜DSML｜tool_calls>\n<｜DSML｜invoke name=\"$TOOL_NAME\">\n<｜DSML｜parameter name=\"$PARAMETER_NAME\" string=\"true|false\">$PARAMETER_VALUE\n...\n\n<｜DSML｜invoke name=\"$TOOL_NAME2\">\n...\n\n\n\nString parameters should be specified as is and set `string=\"true\"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string=\"false\"`.\n\nIf thinking_mode is enabled (triggered by ), you MUST output your complete reasoning inside ... BEFORE any tool calls or final response.\n\nOtherwise, output directly after with tool calls or final response.\n\n### Available Tool Schemas\n\n{\"name\": \"get_weather\", \"description\": \"Get weather for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}\n{\"name\": \"get_time\", \"description\": \"Get local time for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}\n\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/nested_arrays_enums.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.\n\n## Tools\n\nYou have access to a set of tools to help answer the user's question. You can invoke tools by writing a \"<｜DSML｜tool_calls>\" block like the following:\n\n<｜DSML｜tool_calls>\n<｜DSML｜invoke name=\"$TOOL_NAME\">\n<｜DSML｜parameter name=\"$PARAMETER_NAME\" string=\"true|false\">$PARAMETER_VALUE\n...\n\n<｜DSML｜invoke name=\"$TOOL_NAME2\">\n...\n\n\n\nString parameters should be specified as is and set `string=\"true\"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string=\"false\"`.\n\nIf thinking_mode is enabled (triggered by ), you MUST output your complete reasoning inside ... BEFORE any tool calls or final response.\n\nOtherwise, output directly after with tool calls or final response.\n\n### Available Tool Schemas\n\n{\"name\": \"create_ticket\", \"description\": \"Create a support ticket.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"priority\": {\"type\": \"string\", \"enum\": [\"low\", \"medium\", \"high\"]}, \"tags\": {\"type\": \"array\", \"items\": {\"type\": \"string\"}}, \"customer\": {\"type\": \"object\", \"properties\": {\"id\": {\"type\": \"string\"}, \"contacts\": {\"type\": \"array\", \"items\": {\"type\": \"object\", \"properties\": {\"kind\": {\"type\": \"string\", \"enum\": [\"email\", \"phone\"]}, \"value\": {\"type\": \"string\"}}, \"required\": [\"kind\", \"value\"]}}}, \"required\": [\"id\"]}}, \"required\": [\"priority\", \"customer\"]}}\n\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/non_ascii_long_descriptions.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.\n\n## Tools\n\nYou have access to a set of tools to help answer the user's question. You can invoke tools by writing a \"<｜DSML｜tool_calls>\" block like the following:\n\n<｜DSML｜tool_calls>\n<｜DSML｜invoke name=\"$TOOL_NAME\">\n<｜DSML｜parameter name=\"$PARAMETER_NAME\" string=\"true|false\">$PARAMETER_VALUE\n...\n\n<｜DSML｜invoke name=\"$TOOL_NAME2\">\n...\n\n\n\nString parameters should be specified as is and set `string=\"true\"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string=\"false\"`.\n\nIf thinking_mode is enabled (triggered by ), you MUST output your complete reasoning inside ... BEFORE any tool calls or final response.\n\nOtherwise, output directly after with tool calls or final response.\n\n### Available Tool Schemas\n\n{\"name\": \"translate_note\", \"description\": \"Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"text\": {\"type\": \"string\", \"description\": \"Text to translate; may contain unicode like naïve façade résumé.\"}, \"target_language\": {\"type\": \"string\", \"enum\": [\"en\", \"fr\", \"ja\", \"es\"]}}, \"required\": [\"text\", \"target_language\"], \"additionalProperties\": false}}\n\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/nullable_anyof_refs.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.\n\n## Tools\n\nYou have access to a set of tools to help answer the user's question. You can invoke tools by writing a \"<｜DSML｜tool_calls>\" block like the following:\n\n<｜DSML｜tool_calls>\n<｜DSML｜invoke name=\"$TOOL_NAME\">\n<｜DSML｜parameter name=\"$PARAMETER_NAME\" string=\"true|false\">$PARAMETER_VALUE\n...\n\n<｜DSML｜invoke name=\"$TOOL_NAME2\">\n...\n\n\n\nString parameters should be specified as is and set `string=\"true\"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string=\"false\"`.\n\nIf thinking_mode is enabled (triggered by ), you MUST output your complete reasoning inside ... BEFORE any tool calls or final response.\n\nOtherwise, output directly after with tool calls or final response.\n\n### Available Tool Schemas\n\n{\"name\": \"search_inventory\", \"description\": \"Search inventory with advanced filters.\", \"parameters\": {\"$defs\": {\"priceRange\": {\"type\": \"object\", \"properties\": {\"min\": {\"type\": \"number\", \"default\": 0}, \"max\": {\"type\": [\"number\", \"null\"]}}}}, \"type\": \"object\", \"properties\": {\"query\": {\"type\": \"string\"}, \"price\": {\"$ref\": \"#/$defs/priceRange\"}, \"sort\": {\"anyOf\": [{\"const\": \"price\"}, {\"const\": \"rating\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"query\"]}}\n\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/oneof_allof_constraints.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.\n\n## Tools\n\nYou have access to a set of tools to help answer the user's question. You can invoke tools by writing a \"<｜DSML｜tool_calls>\" block like the following:\n\n<｜DSML｜tool_calls>\n<｜DSML｜invoke name=\"$TOOL_NAME\">\n<｜DSML｜parameter name=\"$PARAMETER_NAME\" string=\"true|false\">$PARAMETER_VALUE\n...\n\n<｜DSML｜invoke name=\"$TOOL_NAME2\">\n...\n\n\n\nString parameters should be specified as is and set `string=\"true\"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string=\"false\"`.\n\nIf thinking_mode is enabled (triggered by ), you MUST output your complete reasoning inside ... BEFORE any tool calls or final response.\n\nOtherwise, output directly after with tool calls or final response.\n\n### Available Tool Schemas\n\n{\"name\": \"validate_payload\", \"description\": \"Validate a constrained payload.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"contact\": {\"oneOf\": [{\"type\": \"string\", \"format\": \"email\"}, {\"type\": \"string\", \"pattern\": \"^\\\\+?[0-9-]+$\"}]}, \"scores\": {\"type\": \"array\", \"items\": {\"type\": \"integer\", \"enum\": [1, 2, 3, null]}, \"minItems\": 1, \"maxItems\": 5}, \"profile\": {\"allOf\": [{\"type\": \"object\", \"properties\": {\"name\": {\"type\": \"string\"}}}, {\"type\": \"object\", \"properties\": {\"age\": {\"type\": \"integer\", \"minimum\": 0}}}]}}, \"required\": [\"contact\"], \"additionalProperties\": false}}\n\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/primitive_types.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.\n\n## Tools\n\nYou have access to a set of tools to help answer the user's question. You can invoke tools by writing a \"<｜DSML｜tool_calls>\" block like the following:\n\n<｜DSML｜tool_calls>\n<｜DSML｜invoke name=\"$TOOL_NAME\">\n<｜DSML｜parameter name=\"$PARAMETER_NAME\" string=\"true|false\">$PARAMETER_VALUE\n...\n\n<｜DSML｜invoke name=\"$TOOL_NAME2\">\n...\n\n\n\nString parameters should be specified as is and set `string=\"true\"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string=\"false\"`.\n\nIf thinking_mode is enabled (triggered by ), you MUST output your complete reasoning inside ... BEFORE any tool calls or final response.\n\nOtherwise, output directly after with tool calls or final response.\n\n### Available Tool Schemas\n\n{\"name\": \"book_room\", \"description\": \"Book a hotel room.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"guest_name\": {\"type\": \"string\"}, \"nights\": {\"type\": \"integer\", \"minimum\": 1}, \"max_price\": {\"type\": \"number\"}, \"breakfast\": {\"type\": \"boolean\"}}, \"required\": [\"guest_name\", \"nights\"]}}\n\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/deepseek-ai__DeepSeek-V4-Pro/refs_unions_defaults_constraints.custom_deepseek_dsv4.rendered.txt": "<｜begin▁of▁sentence｜>You are a helpful assistant.\n\n## Tools\n\nYou have access to a set of tools to help answer the user's question. You can invoke tools by writing a \"<｜DSML｜tool_calls>\" block like the following:\n\n<｜DSML｜tool_calls>\n<｜DSML｜invoke name=\"$TOOL_NAME\">\n<｜DSML｜parameter name=\"$PARAMETER_NAME\" string=\"true|false\">$PARAMETER_VALUE\n...\n\n<｜DSML｜invoke name=\"$TOOL_NAME2\">\n...\n\n\n\nString parameters should be specified as is and set `string=\"true\"`. For all other types (numbers, booleans, arrays, objects), pass the value in JSON format and set `string=\"false\"`.\n\nIf thinking_mode is enabled (triggered by ), you MUST output your complete reasoning inside ... BEFORE any tool calls or final response.\n\nOtherwise, output directly after with tool calls or final response.\n\n### Available Tool Schemas\n\n{\"name\": \"plan_shipment\", \"description\": \"Plan a shipment with reusable address and option schemas.\", \"parameters\": {\"$defs\": {\"address\": {\"type\": \"object\", \"description\": \"Postal address.\", \"properties\": {\"street\": {\"type\": \"string\", \"minLength\": 1}, \"postal_code\": {\"type\": \"string\", \"pattern\": \"^[0-9A-Z -]+$\"}}, \"required\": [\"street\", \"postal_code\"]}}, \"type\": \"object\", \"properties\": {\"destination\": {\"$ref\": \"#/$defs/address\"}, \"weight_kg\": {\"type\": \"number\", \"minimum\": 0.1, \"default\": 1.0}, \"delivery_window\": {\"anyOf\": [{\"type\": \"string\", \"description\": \"Named delivery window.\"}, {\"type\": [\"number\", \"null\"], \"description\": \"Unix timestamp or null.\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"destination\", \"weight_kg\"], \"additionalProperties\": false}}\n\nYou MUST strictly follow the above defined tool name and parameter schemas to invoke tool calls.\n<｜User｜>Use the available tools if needed: what is the answer?<｜Assistant｜>", "research/artifacts/rendered/google__gemma-4-31B-it/empty_tools.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/empty_tools.tools_argument.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/function_without_parameters.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/function_without_parameters.tools_argument.rendered.txt": "<|turn>system\n<|tool>declaration:ping{description:<|\"|>Return pong.<|\"|>}\n<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/minimal_string_required.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/minimal_string_required.tools_argument.rendered.txt": "<|turn>system\n<|tool>declaration:lookup_city{description:<|\"|>Look up information about a city.<|\"|>,parameters:{properties:{city:{description:<|\"|>City name.<|\"|>,type:<|\"|>STRING<|\"|>}},required:[<|\"|>city<|\"|>],type:<|\"|>OBJECT<|\"|>}}\n<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "<|turn>system\n<|tool>declaration:get_weather{description:<|\"|>Get weather for a location.<|\"|>,parameters:{properties:{location:{type:<|\"|>STRING<|\"|>}},required:[<|\"|>location<|\"|>],type:<|\"|>OBJECT<|\"|>}}<|tool>declaration:get_time{description:<|\"|>Get local time for a location.<|\"|>,parameters:{properties:{location:{type:<|\"|>STRING<|\"|>}},required:[<|\"|>location<|\"|>],type:<|\"|>OBJECT<|\"|>}}\n<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/nested_arrays_enums.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/nested_arrays_enums.tools_argument.rendered.txt": "<|turn>system\n<|tool>declaration:create_ticket{description:<|\"|>Create a support ticket.<|\"|>,parameters:{properties:{customer:{properties:{contacts:{items:{properties:{kind:{enum:[<|\"|>email<|\"|>,<|\"|>phone<|\"|>],type:<|\"|>STRING<|\"|>},value:{type:<|\"|>STRING<|\"|>}},required:[<|\"|>kind<|\"|>,<|\"|>value<|\"|>],type:<|\"|>OBJECT<|\"|>},type:<|\"|>ARRAY<|\"|>},id:{type:<|\"|>STRING<|\"|>}},required:[<|\"|>id<|\"|>],type:<|\"|>OBJECT<|\"|>},priority:{enum:[<|\"|>low<|\"|>,<|\"|>medium<|\"|>,<|\"|>high<|\"|>],type:<|\"|>STRING<|\"|>},tags:{items:{type:<|\"|>STRING<|\"|>},type:<|\"|>ARRAY<|\"|>}},required:[<|\"|>priority<|\"|>,<|\"|>customer<|\"|>],type:<|\"|>OBJECT<|\"|>}}\n<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/non_ascii_long_descriptions.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/non_ascii_long_descriptions.tools_argument.rendered.txt": "<|turn>system\n<|tool>declaration:translate_note{description:<|\"|>Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.<|\"|>,parameters:{properties:{target_language:{enum:[<|\"|>en<|\"|>,<|\"|>fr<|\"|>,<|\"|>ja<|\"|>,<|\"|>es<|\"|>],type:<|\"|>STRING<|\"|>},text:{description:<|\"|>Text to translate; may contain unicode like naïve façade résumé.<|\"|>,type:<|\"|>STRING<|\"|>}},required:[<|\"|>text<|\"|>,<|\"|>target_language<|\"|>],type:<|\"|>OBJECT<|\"|>}}\n<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/nullable_anyof_refs.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/nullable_anyof_refs.tools_argument.rendered.txt": "<|turn>system\n<|tool>declaration:search_inventory{description:<|\"|>Search inventory with advanced filters.<|\"|>,parameters:{properties:{metadata:{properties:{additionalProperties:{type:<|\"|>STRING<|\"|>}},type:<|\"|>OBJECT<|\"|>},price:{type:<|\"|><|\"|>},query:{type:<|\"|>STRING<|\"|>},sort:{type:<|\"|><|\"|>}},required:[<|\"|>query<|\"|>],type:<|\"|>OBJECT<|\"|>}}\n<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/oneof_allof_constraints.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/oneof_allof_constraints.tools_argument.rendered.txt": "<|turn>system\n<|tool>declaration:validate_payload{description:<|\"|>Validate a constrained payload.<|\"|>,parameters:{properties:{contact:{type:<|\"|><|\"|>},profile:{type:<|\"|><|\"|>},scores:{items:{enum:[1,2,3,None],type:<|\"|>INTEGER<|\"|>},type:<|\"|>ARRAY<|\"|>}},required:[<|\"|>contact<|\"|>],type:<|\"|>OBJECT<|\"|>}}\n<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/primitive_types.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/primitive_types.tools_argument.rendered.txt": "<|turn>system\n<|tool>declaration:book_room{description:<|\"|>Book a hotel room.<|\"|>,parameters:{properties:{breakfast:{type:<|\"|>BOOLEAN<|\"|>},guest_name:{type:<|\"|>STRING<|\"|>},max_price:{type:<|\"|>NUMBER<|\"|>},nights:{type:<|\"|>INTEGER<|\"|>}},required:[<|\"|>guest_name<|\"|>,<|\"|>nights<|\"|>],type:<|\"|>OBJECT<|\"|>}}\n<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/google__gemma-4-31B-it/refs_unions_defaults_constraints.tools_argument.rendered.txt": "<|turn>system\n<|tool>declaration:plan_shipment{description:<|\"|>Plan a shipment with reusable address and option schemas.<|\"|>,parameters:{properties:{delivery_window:{type:<|\"|><|\"|>},destination:{type:<|\"|><|\"|>},metadata:{properties:{additionalProperties:{type:<|\"|>STRING<|\"|>}},type:<|\"|>OBJECT<|\"|>},weight_kg:{type:<|\"|>NUMBER<|\"|>}},required:[<|\"|>destination<|\"|>,<|\"|>weight_kg<|\"|>],type:<|\"|>OBJECT<|\"|>}}\n<|turn>user\nUse the available tools if needed: what is the answer?\n<|turn>model\n<|channel>thought\n", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/empty_tools.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/empty_tools.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/function_without_parameters.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/function_without_parameters.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"type\": \"function\", \"function\": {\"name\": \"ping\", \"description\": \"Return pong.\"}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/minimal_string_required.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/minimal_string_required.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"type\": \"function\", \"function\": {\"name\": \"lookup_city\", \"description\": \"Look up information about a city.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"city\": {\"type\": \"string\", \"description\": \"City name.\"}}, \"required\": [\"city\"]}}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"type\": \"function\", \"function\": {\"name\": \"get_weather\", \"description\": \"Get weather for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}}, {\"type\": \"function\", \"function\": {\"name\": \"get_time\", \"description\": \"Get local time for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nested_arrays_enums.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nested_arrays_enums.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"type\": \"function\", \"function\": {\"name\": \"create_ticket\", \"description\": \"Create a support ticket.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"priority\": {\"type\": \"string\", \"enum\": [\"low\", \"medium\", \"high\"]}, \"tags\": {\"type\": \"array\", \"items\": {\"type\": \"string\"}}, \"customer\": {\"type\": \"object\", \"properties\": {\"id\": {\"type\": \"string\"}, \"contacts\": {\"type\": \"array\", \"items\": {\"type\": \"object\", \"properties\": {\"kind\": {\"type\": \"string\", \"enum\": [\"email\", \"phone\"]}, \"value\": {\"type\": \"string\"}}, \"required\": [\"kind\", \"value\"]}}}, \"required\": [\"id\"]}}, \"required\": [\"priority\", \"customer\"]}}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/non_ascii_long_descriptions.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/non_ascii_long_descriptions.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"type\": \"function\", \"function\": {\"name\": \"translate_note\", \"description\": \"Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"text\": {\"type\": \"string\", \"description\": \"Text to translate; may contain unicode like naïve façade résumé.\"}, \"target_language\": {\"type\": \"string\", \"enum\": [\"en\", \"fr\", \"ja\", \"es\"]}}, \"required\": [\"text\", \"target_language\"], \"additionalProperties\": false}}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nullable_anyof_refs.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/nullable_anyof_refs.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"type\": \"function\", \"function\": {\"name\": \"search_inventory\", \"description\": \"Search inventory with advanced filters.\", \"parameters\": {\"$defs\": {\"priceRange\": {\"type\": \"object\", \"properties\": {\"min\": {\"type\": \"number\", \"default\": 0}, \"max\": {\"type\": [\"number\", \"null\"]}}}}, \"type\": \"object\", \"properties\": {\"query\": {\"type\": \"string\"}, \"price\": {\"$ref\": \"#/$defs/priceRange\"}, \"sort\": {\"anyOf\": [{\"const\": \"price\"}, {\"const\": \"rating\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"query\"]}}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/oneof_allof_constraints.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/oneof_allof_constraints.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"type\": \"function\", \"function\": {\"name\": \"validate_payload\", \"description\": \"Validate a constrained payload.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"contact\": {\"oneOf\": [{\"type\": \"string\", \"format\": \"email\"}, {\"type\": \"string\", \"pattern\": \"^\\\\+?[0-9-]+$\"}]}, \"scores\": {\"type\": \"array\", \"items\": {\"type\": \"integer\", \"enum\": [1, 2, 3, null]}, \"minItems\": 1, \"maxItems\": 5}, \"profile\": {\"allOf\": [{\"type\": \"object\", \"properties\": {\"name\": {\"type\": \"string\"}}}, {\"type\": \"object\", \"properties\": {\"age\": {\"type\": \"integer\", \"minimum\": 0}}}]}}, \"required\": [\"contact\"], \"additionalProperties\": false}}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/primitive_types.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/primitive_types.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"type\": \"function\", \"function\": {\"name\": \"book_room\", \"description\": \"Book a hotel room.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"guest_name\": {\"type\": \"string\"}, \"nights\": {\"type\": \"integer\", \"minimum\": 1}, \"max_price\": {\"type\": \"number\"}, \"breakfast\": {\"type\": \"boolean\"}}, \"required\": [\"guest_name\", \"nights\"]}}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/raw_json_schema_parameters_only.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"name\": \"raw_schema_probe\", \"description\": \"Probe non-wrapper compatibility.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"q\": {\"type\": \"string\", \"pattern\": \"^[A-Z].+\"}}, \"required\": [\"q\"], \"additionalProperties\": false}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/mistralai__Mistral-Medium-3.5-128B/refs_unions_defaults_constraints.tools_argument.rendered.txt": "[SYSTEM_PROMPT]You are Mistral Medium 3.5, a Large Language Model (LLM) created by Mistral AI, a French startup headquartered in Paris.\nYou are an intelligent conversational assistant powering an AI assistant called Le Chat.\nYour knowledge base was last updated on Friday, November 1, 2024.\nThe current date is 29-04-2026.\n\n# GENERAL GUIDELINES\n\n- Accurately answer the user's question.\n- For uncertain information or when the user's request requires up-to-date or specific data, use the available tools to fetch the information.\n- Be very attentive to dates, always try to resolve dates (e.g. \"yesterday\" is 28-04-2026) and when asked about information at specific dates, discard information that is at another date.\n\n# WEB BROWSING INSTRUCTIONS\n\nYou cannot perform any web search or access internet to open URLs, links etc without dedicated tools.\n\n# MULTI-MODAL INSTRUCTIONS\n\n- You have the ability to read images.\n- You cannot read audio nor videos.\n- You cannot generate images without dedicated tools.\n\n# TOOL CALLING INSTRUCTIONS\n\nYou may have access to tools that you can use to fetch information or perform actions. You must use these tools in the following situations:\n\n1. When the request requires up-to-date information.\n2. When the request requires specific data that you do not have in your knowledge base.\n3. When the request involves actions that you cannot perform without tools.\n\nAlways prioritize using tools to provide the most accurate and helpful response.[/SYSTEM_PROMPT][AVAILABLE_TOOLS][{\"type\": \"function\", \"function\": {\"name\": \"plan_shipment\", \"description\": \"Plan a shipment with reusable address and option schemas.\", \"parameters\": {\"$defs\": {\"address\": {\"type\": \"object\", \"description\": \"Postal address.\", \"properties\": {\"street\": {\"type\": \"string\", \"minLength\": 1}, \"postal_code\": {\"type\": \"string\", \"pattern\": \"^[0-9A-Z -]+$\"}}, \"required\": [\"street\", \"postal_code\"]}}, \"type\": \"object\", \"properties\": {\"destination\": {\"$ref\": \"#/$defs/address\"}, \"weight_kg\": {\"type\": \"number\", \"minimum\": 0.1, \"default\": 1.0}, \"delivery_window\": {\"anyOf\": [{\"type\": \"string\", \"description\": \"Named delivery window.\"}, {\"type\": [\"number\", \"null\"], \"description\": \"Unix timestamp or null.\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"destination\", \"weight_kg\"], \"additionalProperties\": false}}}][/AVAILABLE_TOOLS][MODEL_SETTINGS]{\"reasoning_effort\": \"none\"}[/MODEL_SETTINGS][INST]Use the available tools if needed: what is the answer?[/INST]", "research/artifacts/rendered/moonshotai__Kimi-K2.5/empty_tools.custom_kimi_ts.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/empty_tools.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/empty_tools.tools_argument.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/function_without_parameters.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Return pong.\ntype ping = (_: {}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/function_without_parameters.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/minimal_string_required.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Look up information about a city.\ntype lookup_city = (_: {\n // City name.\n city: string\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/minimal_string_required.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/multiple_tools_parallel_candidate.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Get weather for a location.\ntype get_weather = (_: {\n location: string\n}) => any;\n// Get local time for a location.\ntype get_time = (_: {\n location: string\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/nested_arrays_enums.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Create a support ticket.\ntype create_ticket = (_: {\n customer: {\n id: string,\n contacts?: Array<{\n kind: \"email\" | \"phone\",\n value: string\n }>\n },\n priority: \"low\" | \"medium\" | \"high\",\n tags?: Array\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/nested_arrays_enums.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/non_ascii_long_descriptions.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\ntype translate_note = (_: {\n target_language: \"en\" | \"fr\" | \"ja\" | \"es\",\n // Text to translate; may contain unicode like naïve façade résumé.\n text: string,\n [k: string]: never\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/non_ascii_long_descriptions.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/nullable_anyof_refs.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/oneof_allof_constraints.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/primitive_types.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Book a hotel room.\ntype book_room = (_: {\n guest_name: string,\n // minimum: 1\n nights: number,\n breakfast?: boolean,\n max_price?: number\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/primitive_types.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/raw_json_schema_parameters_only.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|>[{\"name\": \"raw_schema_probe\", \"description\": \"Probe non-wrapper compatibility.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"q\": {\"type\": \"string\", \"pattern\": \"^[A-Z].+\"}}, \"required\": [\"q\"], \"additionalProperties\": false}}]<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/refs_unions_defaults_constraints.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Postal address.\ninterface address {\n // pattern: ^[0-9A-Z -]+$\n postal_code: string,\n // minLength: 1\n street: string\n}\n// Plan a shipment with reusable address and option schemas.\ntype plan_shipment = (_: {\n destination: address,\n // minimum: 0.1\n // Default: 1.0\n weight_kg: number,\n delivery_window?: string | number | null,\n metadata?: {\n [k: string]: string\n },\n [k: string]: never\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.5/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/empty_tools.custom_kimi_ts.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/empty_tools.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/empty_tools.tools_argument.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/function_without_parameters.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Return pong.\ntype ping = (_: {}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/function_without_parameters.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/minimal_string_required.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Look up information about a city.\ntype lookup_city = (_: {\n // City name.\n city: string\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/minimal_string_required.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/multiple_tools_parallel_candidate.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Get weather for a location.\ntype get_weather = (_: {\n location: string\n}) => any;\n// Get local time for a location.\ntype get_time = (_: {\n location: string\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/nested_arrays_enums.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Create a support ticket.\ntype create_ticket = (_: {\n customer: {\n id: string,\n contacts?: Array<{\n kind: \"email\" | \"phone\",\n value: string\n }>\n },\n priority: \"low\" | \"medium\" | \"high\",\n tags?: Array\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/nested_arrays_enums.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/non_ascii_long_descriptions.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\ntype translate_note = (_: {\n target_language: \"en\" | \"fr\" | \"ja\" | \"es\",\n // Text to translate; may contain unicode like naïve façade résumé.\n text: string,\n [k: string]: never\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/non_ascii_long_descriptions.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/nullable_anyof_refs.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/oneof_allof_constraints.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/primitive_types.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Book a hotel room.\ntype book_room = (_: {\n guest_name: string,\n // minimum: 1\n nights: number,\n breakfast?: boolean,\n max_price?: number\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/primitive_types.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/raw_json_schema_parameters_only.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|>[{\"name\": \"raw_schema_probe\", \"description\": \"Probe non-wrapper compatibility.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"q\": {\"type\": \"string\", \"pattern\": \"^[A-Z].+\"}}, \"required\": [\"q\"], \"additionalProperties\": false}}]<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/refs_unions_defaults_constraints.custom_kimi_ts.rendered.txt": "<|im_system|>tool_declare<|im_middle|># Tools\n\n## functions\nnamespace functions {\n// Postal address.\ninterface address {\n // pattern: ^[0-9A-Z -]+$\n postal_code: string,\n // minLength: 1\n street: string\n}\n// Plan a shipment with reusable address and option schemas.\ntype plan_shipment = (_: {\n destination: address,\n // minimum: 0.1\n // Default: 1.0\n weight_kg: number,\n delivery_window?: string | number | null,\n metadata?: {\n [k: string]: string\n },\n [k: string]: never\n}) => any;\n}\n<|im_end|><|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/moonshotai__Kimi-K2.6/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "<|im_user|>user<|im_middle|>Use the available tools if needed: what is the answer?<|im_end|><|im_assistant|>assistant<|im_middle|>", "research/artifacts/rendered/openai__gpt-oss-120b/empty_tools.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/empty_tools.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/function_without_parameters.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/function_without_parameters.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.\nCalls to these tools must go to the commentary channel: 'functions'.<|end|><|start|>developer<|message|># Tools\n\n## functions\n\nnamespace functions {\n\n// Return pong.\ntype ping = () => any;\n\n} // namespace functions<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/minimal_string_required.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/minimal_string_required.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.\nCalls to these tools must go to the commentary channel: 'functions'.<|end|><|start|>developer<|message|># Tools\n\n## functions\n\nnamespace functions {\n\n// Look up information about a city.\ntype lookup_city = (_: {\n// City name.\ncity: string,\n}) => any;\n\n} // namespace functions<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.\nCalls to these tools must go to the commentary channel: 'functions'.<|end|><|start|>developer<|message|># Tools\n\n## functions\n\nnamespace functions {\n\n// Get weather for a location.\ntype get_weather = (_: {\nlocation: string,\n}) => any;\n\n// Get local time for a location.\ntype get_time = (_: {\nlocation: string,\n}) => any;\n\n} // namespace functions<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/nested_arrays_enums.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/nested_arrays_enums.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.\nCalls to these tools must go to the commentary channel: 'functions'.<|end|><|start|>developer<|message|># Tools\n\n## functions\n\nnamespace functions {\n\n// Create a support ticket.\ntype create_ticket = (_: {\npriority: \"low\" | \"medium\" | \"high\",\ntags?: string[],\ncustomer: {\nid: \n string, contacts?: \n any[]},\n}) => any;\n\n} // namespace functions<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/non_ascii_long_descriptions.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/non_ascii_long_descriptions.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.\nCalls to these tools must go to the commentary channel: 'functions'.<|end|><|start|>developer<|message|># Tools\n\n## functions\n\nnamespace functions {\n\n// Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\ntype translate_note = (_: {\n// Text to translate; may contain unicode like naïve façade résumé.\ntext: string,\ntarget_language: \"en\" | \"fr\" | \"ja\" | \"es\",\n}) => any;\n\n} // namespace functions<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/nullable_anyof_refs.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/nullable_anyof_refs.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.\nCalls to these tools must go to the commentary channel: 'functions'.<|end|><|start|>developer<|message|># Tools\n\n## functions\n\nnamespace functions {\n\n// Search inventory with advanced filters.\ntype search_inventory = (_: {\nquery: string,\nprice?: any,\nsort?: any,\nmetadata?: object,\n}) => any;\n\n} // namespace functions<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/oneof_allof_constraints.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/oneof_allof_constraints.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.\nCalls to these tools must go to the commentary channel: 'functions'.<|end|><|start|>developer<|message|># Tools\n\n## functions\n\nnamespace functions {\n\n// Validate a constrained payload.\ntype validate_payload = (_: {\ncontact: string | \nstring,\nscores?: number[],\nprofile?: any,\n}) => any;\n\n} // namespace functions<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/primitive_types.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/primitive_types.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.\nCalls to these tools must go to the commentary channel: 'functions'.<|end|><|start|>developer<|message|># Tools\n\n## functions\n\nnamespace functions {\n\n// Book a hotel room.\ntype book_room = (_: {\nguest_name: string,\nnights: number,\nmax_price?: number,\nbreakfast?: boolean,\n}) => any;\n\n} // namespace functions<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/refs_unions_defaults_constraints.tools_argument.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.\nCalls to these tools must go to the commentary channel: 'functions'.<|end|><|start|>developer<|message|># Tools\n\n## functions\n\nnamespace functions {\n\n// Plan a shipment with reusable address and option schemas.\ntype plan_shipment = (_: {\ndestination: any,\nweight_kg: number, // default: 1.0,\ndelivery_window?: any,\nmetadata?: object,\n}) => any;\n\n} // namespace functions<|end|><|start|>user<|message|>Use the available tools if needed: what is the answer?<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_browser.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Tools\n\n## browser\n\n// Tool for browsing.\n// The `cursor` appears in brackets before each browsing display: `[{cursor}]`.\n// Cite information from the tool using the following format:\n// `【{cursor}†L{line_start}(-L{line_end})?】`, for example: `【6†L9-L11】` or `【8†L3】`.\n// Do not quote more than 10 words directly from the tool output.\n// sources=web (default: web)\nnamespace browser {\n\n// Searches for information related to `query` and displays `topn` results.\ntype search = (_: {\nquery: string,\ntopn?: number, // default: 10\nsource?: string,\n}) => any;\n\n// Opens the link `id` from the page indicated by `cursor` starting at line number `loc`, showing `num_lines` lines.\n// Valid link ids are displayed with the formatting: `【{id}†.*】`.\n// If `cursor` is not provided, the most recent page is implied.\n// If `id` is a string, it is treated as a fully qualified URL associated with `source`.\n// If `loc` is not provided, the viewport will be positioned at the beginning of the document or centered on the most relevant passage, if available.\n// Use this function without `id` to scroll to a new location of an opened page.\ntype open = (_: {\nid?: number | string, // default: -1\ncursor?: number, // default: -1\nloc?: number, // default: -1\nnum_lines?: number, // default: -1\nview_source?: boolean, // default: false\nsource?: string,\n}) => any;\n\n// Finds exact matches of `pattern` in the current page, or the page given by `cursor`.\ntype find = (_: {\npattern: string,\ncursor?: number, // default: -1\n}) => any;\n\n} // namespace browser\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Solve 2+2. Use special tools if appropriate.<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_browser_python.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Tools\n\n## browser\n\n// Tool for browsing.\n// The `cursor` appears in brackets before each browsing display: `[{cursor}]`.\n// Cite information from the tool using the following format:\n// `【{cursor}†L{line_start}(-L{line_end})?】`, for example: `【6†L9-L11】` or `【8†L3】`.\n// Do not quote more than 10 words directly from the tool output.\n// sources=web (default: web)\nnamespace browser {\n\n// Searches for information related to `query` and displays `topn` results.\ntype search = (_: {\nquery: string,\ntopn?: number, // default: 10\nsource?: string,\n}) => any;\n\n// Opens the link `id` from the page indicated by `cursor` starting at line number `loc`, showing `num_lines` lines.\n// Valid link ids are displayed with the formatting: `【{id}†.*】`.\n// If `cursor` is not provided, the most recent page is implied.\n// If `id` is a string, it is treated as a fully qualified URL associated with `source`.\n// If `loc` is not provided, the viewport will be positioned at the beginning of the document or centered on the most relevant passage, if available.\n// Use this function without `id` to scroll to a new location of an opened page.\ntype open = (_: {\nid?: number | string, // default: -1\ncursor?: number, // default: -1\nloc?: number, // default: -1\nnum_lines?: number, // default: -1\nview_source?: boolean, // default: false\nsource?: string,\n}) => any;\n\n// Finds exact matches of `pattern` in the current page, or the page given by `cursor`.\ntype find = (_: {\npattern: string,\ncursor?: number, // default: -1\n}) => any;\n\n} // namespace browser\n\n## python\n\nUse this tool to execute Python code in your chain of thought. The code will not be shown to the user. This tool should be used for internal reasoning, but not for code that is intended to be visible to the user (e.g. when creating plots, tables, or files).\n\nWhen you send a message containing Python code to python, it will be executed in a stateful Jupyter notebook environment. python will respond with the output of the execution or time out after 120.0 seconds. The drive at '/mnt/data' can be used to save and persist user files. Internet access for this session is UNKNOWN. Depends on the cluster.\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Solve 2+2. Use special tools if appropriate.<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_medium.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Tools\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Solve 2+2. Use special tools if appropriate.<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_model_identity.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Tools\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Solve 2+2. Use special tools if appropriate.<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_python.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Tools\n\n## python\n\nUse this tool to execute Python code in your chain of thought. The code will not be shown to the user. This tool should be used for internal reasoning, but not for code that is intended to be visible to the user (e.g. when creating plots, tables, or files).\n\nWhen you send a message containing Python code to python, it will be executed in a stateful Jupyter notebook environment. python will respond with the output of the execution or time out after 120.0 seconds. The drive at '/mnt/data' can be used to save and persist user files. Internet access for this session is UNKNOWN. Depends on the cluster.\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Solve 2+2. Use special tools if appropriate.<|end|><|start|>assistant", "research/artifacts/rendered/openai__gpt-oss-120b/special_tool_modes/builtin_reasoning_effort.rendered.txt": "<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.\nKnowledge cutoff: 2024-06\nCurrent date: 2026-05-08\n\nReasoning: medium\n\n# Tools\n\n# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|><|start|>user<|message|>Solve 2+2. Use special tools if appropriate.<|end|><|start|>assistant", "research/artifacts/rendered/zai-org__GLM-5/empty_tools.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/empty_tools.tools_argument.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/function_without_parameters.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/function_without_parameters.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"type\": \"function\", \"function\": {\"name\": \"ping\", \"description\": \"Return pong.\"}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/minimal_string_required.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/minimal_string_required.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"type\": \"function\", \"function\": {\"name\": \"lookup_city\", \"description\": \"Look up information about a city.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"city\": {\"type\": \"string\", \"description\": \"City name.\"}}, \"required\": [\"city\"]}}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"type\": \"function\", \"function\": {\"name\": \"get_weather\", \"description\": \"Get weather for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}}\n{\"type\": \"function\", \"function\": {\"name\": \"get_time\", \"description\": \"Get local time for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/nested_arrays_enums.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/nested_arrays_enums.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"type\": \"function\", \"function\": {\"name\": \"create_ticket\", \"description\": \"Create a support ticket.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"priority\": {\"type\": \"string\", \"enum\": [\"low\", \"medium\", \"high\"]}, \"tags\": {\"type\": \"array\", \"items\": {\"type\": \"string\"}}, \"customer\": {\"type\": \"object\", \"properties\": {\"id\": {\"type\": \"string\"}, \"contacts\": {\"type\": \"array\", \"items\": {\"type\": \"object\", \"properties\": {\"kind\": {\"type\": \"string\", \"enum\": [\"email\", \"phone\"]}, \"value\": {\"type\": \"string\"}}, \"required\": [\"kind\", \"value\"]}}}, \"required\": [\"id\"]}}, \"required\": [\"priority\", \"customer\"]}}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/non_ascii_long_descriptions.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/non_ascii_long_descriptions.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"type\": \"function\", \"function\": {\"name\": \"translate_note\", \"description\": \"Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"text\": {\"type\": \"string\", \"description\": \"Text to translate; may contain unicode like naïve façade résumé.\"}, \"target_language\": {\"type\": \"string\", \"enum\": [\"en\", \"fr\", \"ja\", \"es\"]}}, \"required\": [\"text\", \"target_language\"], \"additionalProperties\": false}}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/nullable_anyof_refs.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/nullable_anyof_refs.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"type\": \"function\", \"function\": {\"name\": \"search_inventory\", \"description\": \"Search inventory with advanced filters.\", \"parameters\": {\"$defs\": {\"priceRange\": {\"type\": \"object\", \"properties\": {\"min\": {\"type\": \"number\", \"default\": 0}, \"max\": {\"type\": [\"number\", \"null\"]}}}}, \"type\": \"object\", \"properties\": {\"query\": {\"type\": \"string\"}, \"price\": {\"$ref\": \"#/$defs/priceRange\"}, \"sort\": {\"anyOf\": [{\"const\": \"price\"}, {\"const\": \"rating\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"query\"]}}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/oneof_allof_constraints.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/oneof_allof_constraints.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"type\": \"function\", \"function\": {\"name\": \"validate_payload\", \"description\": \"Validate a constrained payload.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"contact\": {\"oneOf\": [{\"type\": \"string\", \"format\": \"email\"}, {\"type\": \"string\", \"pattern\": \"^\\\\+?[0-9-]+$\"}]}, \"scores\": {\"type\": \"array\", \"items\": {\"type\": \"integer\", \"enum\": [1, 2, 3, null]}, \"minItems\": 1, \"maxItems\": 5}, \"profile\": {\"allOf\": [{\"type\": \"object\", \"properties\": {\"name\": {\"type\": \"string\"}}}, {\"type\": \"object\", \"properties\": {\"age\": {\"type\": \"integer\", \"minimum\": 0}}}]}}, \"required\": [\"contact\"], \"additionalProperties\": false}}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/primitive_types.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/primitive_types.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"type\": \"function\", \"function\": {\"name\": \"book_room\", \"description\": \"Book a hotel room.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"guest_name\": {\"type\": \"string\"}, \"nights\": {\"type\": \"integer\", \"minimum\": 1}, \"max_price\": {\"type\": \"number\"}, \"breakfast\": {\"type\": \"boolean\"}}, \"required\": [\"guest_name\", \"nights\"]}}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/raw_json_schema_parameters_only.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"raw_schema_probe\", \"description\": \"Probe non-wrapper compatibility.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"q\": {\"type\": \"string\", \"pattern\": \"^[A-Z].+\"}}, \"required\": [\"q\"], \"additionalProperties\": false}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5/refs_unions_defaults_constraints.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"type\": \"function\", \"function\": {\"name\": \"plan_shipment\", \"description\": \"Plan a shipment with reusable address and option schemas.\", \"parameters\": {\"$defs\": {\"address\": {\"type\": \"object\", \"description\": \"Postal address.\", \"properties\": {\"street\": {\"type\": \"string\", \"minLength\": 1}, \"postal_code\": {\"type\": \"string\", \"pattern\": \"^[0-9A-Z -]+$\"}}, \"required\": [\"street\", \"postal_code\"]}}, \"type\": \"object\", \"properties\": {\"destination\": {\"$ref\": \"#/$defs/address\"}, \"weight_kg\": {\"type\": \"number\", \"minimum\": 0.1, \"default\": 1.0}, \"delivery_window\": {\"anyOf\": [{\"type\": \"string\", \"description\": \"Named delivery window.\"}, {\"type\": [\"number\", \"null\"], \"description\": \"Unix timestamp or null.\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"destination\", \"weight_kg\"], \"additionalProperties\": false}}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/empty_tools.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/empty_tools.tools_argument.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/function_without_parameters.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/function_without_parameters.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"ping\", \"description\": \"Return pong.\"}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/minimal_string_required.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/minimal_string_required.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"lookup_city\", \"description\": \"Look up information about a city.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"city\": {\"type\": \"string\", \"description\": \"City name.\"}}, \"required\": [\"city\"]}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/multiple_tools_parallel_candidate.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/multiple_tools_parallel_candidate.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"get_weather\", \"description\": \"Get weather for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}\n{\"name\": \"get_time\", \"description\": \"Get local time for a location.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"location\": {\"type\": \"string\"}}, \"required\": [\"location\"]}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/nested_arrays_enums.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/nested_arrays_enums.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"create_ticket\", \"description\": \"Create a support ticket.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"priority\": {\"type\": \"string\", \"enum\": [\"low\", \"medium\", \"high\"]}, \"tags\": {\"type\": \"array\", \"items\": {\"type\": \"string\"}}, \"customer\": {\"type\": \"object\", \"properties\": {\"id\": {\"type\": \"string\"}, \"contacts\": {\"type\": \"array\", \"items\": {\"type\": \"object\", \"properties\": {\"kind\": {\"type\": \"string\", \"enum\": [\"email\", \"phone\"]}, \"value\": {\"type\": \"string\"}}, \"required\": [\"kind\", \"value\"]}}}, \"required\": [\"id\"]}}, \"required\": [\"priority\", \"customer\"]}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/non_ascii_long_descriptions.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/non_ascii_long_descriptions.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"translate_note\", \"description\": \"Translate a user note. Handles café, 東京, emojis like 🚀, and long natural-language guidance that may stress template formatting.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"text\": {\"type\": \"string\", \"description\": \"Text to translate; may contain unicode like naïve façade résumé.\"}, \"target_language\": {\"type\": \"string\", \"enum\": [\"en\", \"fr\", \"ja\", \"es\"]}}, \"required\": [\"text\", \"target_language\"], \"additionalProperties\": false}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/nullable_anyof_refs.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/nullable_anyof_refs.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"search_inventory\", \"description\": \"Search inventory with advanced filters.\", \"parameters\": {\"$defs\": {\"priceRange\": {\"type\": \"object\", \"properties\": {\"min\": {\"type\": \"number\", \"default\": 0}, \"max\": {\"type\": [\"number\", \"null\"]}}}}, \"type\": \"object\", \"properties\": {\"query\": {\"type\": \"string\"}, \"price\": {\"$ref\": \"#/$defs/priceRange\"}, \"sort\": {\"anyOf\": [{\"const\": \"price\"}, {\"const\": \"rating\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"query\"]}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/oneof_allof_constraints.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/oneof_allof_constraints.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"validate_payload\", \"description\": \"Validate a constrained payload.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"contact\": {\"oneOf\": [{\"type\": \"string\", \"format\": \"email\"}, {\"type\": \"string\", \"pattern\": \"^\\\\+?[0-9-]+$\"}]}, \"scores\": {\"type\": \"array\", \"items\": {\"type\": \"integer\", \"enum\": [1, 2, 3, null]}, \"minItems\": 1, \"maxItems\": 5}, \"profile\": {\"allOf\": [{\"type\": \"object\", \"properties\": {\"name\": {\"type\": \"string\"}}}, {\"type\": \"object\", \"properties\": {\"age\": {\"type\": \"integer\", \"minimum\": 0}}}]}}, \"required\": [\"contact\"], \"additionalProperties\": false}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/primitive_types.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/primitive_types.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"book_room\", \"description\": \"Book a hotel room.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"guest_name\": {\"type\": \"string\"}, \"nights\": {\"type\": \"integer\", \"minimum\": 1}, \"max_price\": {\"type\": \"number\"}, \"breakfast\": {\"type\": \"boolean\"}}, \"required\": [\"guest_name\", \"nights\"]}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/raw_json_schema_parameters_only.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/raw_json_schema_parameters_only.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"raw_schema_probe\", \"description\": \"Probe non-wrapper compatibility.\", \"parameters\": {\"type\": \"object\", \"properties\": {\"q\": {\"type\": \"string\", \"pattern\": \"^[A-Z].+\"}}, \"required\": [\"q\"], \"additionalProperties\": false}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/refs_unions_defaults_constraints.message_functions_json.rendered.txt": "[gMASK]<|user|>Use the available tools if needed: what is the answer?<|assistant|>", "research/artifacts/rendered/zai-org__GLM-5.1/refs_unions_defaults_constraints.tools_argument.rendered.txt": "[gMASK]<|system|>\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n\n{\"name\": \"plan_shipment\", \"description\": \"Plan a shipment with reusable address and option schemas.\", \"parameters\": {\"$defs\": {\"address\": {\"type\": \"object\", \"description\": \"Postal address.\", \"properties\": {\"street\": {\"type\": \"string\", \"minLength\": 1}, \"postal_code\": {\"type\": \"string\", \"pattern\": \"^[0-9A-Z -]+$\"}}, \"required\": [\"street\", \"postal_code\"]}}, \"type\": \"object\", \"properties\": {\"destination\": {\"$ref\": \"#/$defs/address\"}, \"weight_kg\": {\"type\": \"number\", \"minimum\": 0.1, \"default\": 1.0}, \"delivery_window\": {\"anyOf\": [{\"type\": \"string\", \"description\": \"Named delivery window.\"}, {\"type\": [\"number\", \"null\"], \"description\": \"Unix timestamp or null.\"}]}, \"metadata\": {\"type\": \"object\", \"additionalProperties\": {\"type\": \"string\"}}}, \"required\": [\"destination\", \"weight_kg\"], \"additionalProperties\": false}}\n\n\nFor each function call, output the function name and arguments within the following XML format:\n{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...<|user|>Use the available tools if needed: what is the answer?<|assistant|>" } };