Add dataset card and release metadata

Add the reviewed English dataset card, CC BY 4.0 terms, citation metadata, manifest, and QA report for v1.0.0.

CITATION.cff

@@ -0,0 +1,23 @@
+cff-version: 1.2.0
+message: "If you use this benchmark, cite the article, dataset release, frozen protocol, and archival record."
+title: "DeepSeek 1M Context Benchmark: Retrieval Accuracy, Latency, and Cost"
+type: dataset
+authors:
+  - name: "Chat Deep AI"
+    website: "https://chat-deep.ai/"
+version: 1.0.0
+date-released: 2026-08-07
+doi: 10.5281/zenodo.21838863
+repository-code: "https://github.com/chatdeepai/deepseek-1m-context-benchmark"
+url: "https://doi.org/10.5281/zenodo.21838863"
+license: CC-BY-4.0
+keywords:
+  - DeepSeek
+  - long context
+  - retrieval benchmark
+  - latency
+  - reproducibility
+abstract: >-
+  A reproducible benchmark of DeepSeek V4 Flash and Pro across deterministic
+  long-context retrieval and synthesis tasks, with sanitized attempt-level
+  data, strict grading, and a frozen methodology.

DATA-DICTIONARY.md

@@ -0,0 +1,59 @@
+# Public data dictionary
+
+The case CSVs are reconstructed through an explicit allowlist. Empty cells mean the provider or adapter did not supply a value; they are not silently imputed.
+
+| Field | Meaning |
+|---|---|
+| `protocol_id` | Frozen benchmark protocol identifier. |
+| `protocol_version` | Frozen protocol semantic version. |
+| `analysis_role` | pilot_excluded, primary_accuracy, or india_latency_validation. |
+| `include_in_primary_accuracy_denominators` | True only for the 288 U.S. primary rows. |
+| `region_vantage` | AWS client network vantage used to send the request; not provider hosting or user location. |
+| `case_id` | Deterministic effective case identifier; not a provider request ID. |
+| `public_case_uid` | Release-specific hash of protocol, role, region, and case_id; unique across all 344 public rows. |
+| `model_id` | Exact requested DeepSeek API model ID. |
+| `family_id` | Frozen objective task family. |
+| `tier_id` | Frozen provider-counted prompt-token tier. |
+| `position_id` | Frozen beginning, middle, or end target position label. |
+| `repeat` | Deterministic fixture repeat number. |
+| `started_at` | UTC request start timestamp. |
+| `finished_at` | UTC terminal timestamp. |
+| `http_status` | Observed HTTP status when available. |
+| `transport_complete` | True only when the HTTP stream completed under the frozen adapter contract. |
+| `finish_reason` | Provider-returned completion finish reason. |
+| `returned_model` | Provider-returned model identifier. |
+| `system_fingerprint` | Provider-returned backend fingerprint when supplied. |
+| `time_to_first_sse_event_ms` | Milliseconds from request start to the first SSE event. |
+| `time_to_first_reasoning_token_ms` | Milliseconds to the first reasoning token; normally blank because thinking was disabled. |
+| `time_to_first_answer_token_ms` | Milliseconds from request start to the first answer token. |
+| `time_to_end_ms` | Milliseconds from request start to stream termination. |
+| `prompt_tokens` | Provider-returned prompt-token count. |
+| `completion_tokens` | Provider-returned completion-token count. |
+| `total_tokens` | Provider-returned total-token count. |
+| `prompt_cache_hit_tokens` | Provider-returned prompt cache-hit tokens. |
+| `prompt_cache_miss_tokens` | Provider-returned prompt cache-miss tokens. |
+| `reasoning_tokens` | Provider-returned reasoning-token count when supplied. |
+| `valid_json` | Whether the final answer was one parseable JSON object. |
+| `exact_keys` | Whether the parsed object had exactly the family-specific key set. |
+| `exact_match` | Frozen primary outcome: exact keys, types, and values with no surrounding prose. |
+| `field_accuracy` | Matched required fields divided by total required fields for the case. |
+| `matched_fields` | Number of exact required fields. |
+| `total_fields` | Number of required fields for the case family. |
+| `grader_errors` | Pipe-delimited public-safe grader error labels. |
+| `prompt_tier_calibration_passed` | Whether provider-returned prompt_tokens fell inside the frozen accepted tier. |
+| `thinking_disabled_violation` | True if reasoning content appeared despite the frozen non-thinking request. |
+| `sse_event_count` | Count of captured SSE events. |
+| `sse_parse_error_count` | Count of SSE parse errors. |
+| `sse_done_received` | Whether the terminal SSE marker was captured. |
+| `stream_error_class` | Public-safe stream error class, if any. |
+| `error_class` | Public-safe adapter error class, if any. |
+| `observed_provider_cost_usd_cache_miss_upper_bound` | Dated cache-miss upper-bound cost calculated from returned token counts. |
+| `generator_bundle_sha256` | SHA-256 of the frozen generator bundle. |
+| `base_fixture_sha256` | SHA-256 of the unsalted deterministic fixture. |
+| `execution_salt_sha256` | SHA-256 of the role-and-region execution salt. |
+| `salted_user_prompt_sha256` | SHA-256 of the sent salted user prompt. |
+| `expected_sha256` | SHA-256 of the objective expected JSON value. |
+| `request_sha256` | SHA-256 of the canonical sent request body. |
+| `raw_response_sha256` | SHA-256 of the immutable raw provider response bytes; raw bytes are not public. |
+
+The sanitized validator export exposes aggregate field counts but not per-field expected and observed values. No per-field row dataset is synthesized.

LICENSE-DATA.txt

@@ -0,0 +1 @@
+Dataset and original charts are licensed under CC BY 4.0: https://creativecommons.org/licenses/by/4.0/

MANIFEST.json

@@ -0,0 +1,252 @@
+{
+  "schema_version": "1.0.0",
+  "bundle_name": "deepseek-1m-context-benchmark-huggingface-dataset-v1.0.0",
+  "created_on": "2026-08-07",
+  "status": "local_ready_not_uploaded",
+  "external_upload_performed": false,
+  "source": {
+    "repository": "https://github.com/chatdeepai/deepseek-1m-context-benchmark",
+    "release": "https://github.com/chatdeepai/deepseek-1m-context-benchmark/releases/tag/v1.0.0",
+    "release_version": "1.0.0",
+    "release_tag": "v1.0.0",
+    "release_commit": "06162a211ad9929646d53d60b97c31b758192d09",
+    "protocol_id": "deepseek-v4-long-context-retrieval-v1.1.0",
+    "article": "https://chat-deep.ai/research/deepseek-1m-context-benchmark/",
+    "trusted_source_archive_sha256": "38337e1213e2cf7ba7de8f21703e11772eddb71326361a66f27f7a8d0c114dcb"
+  },
+  "doi": "10.5281/zenodo.21838863",
+  "doi_url": "https://doi.org/10.5281/zenodo.21838863",
+  "manifest_scope": "Every bundle file except MANIFEST.json itself.",
+  "file_count_in_manifest": 25,
+  "expected_bundle_file_count_including_manifest": 26,
+  "byte_preserved_source_files": 22,
+  "authored_files_in_manifest_scope": 3,
+  "authored_bundle_files_including_manifest": 4,
+  "files": [
+    {
+      "path": "data/all-attempts.csv",
+      "bytes": 332682,
+      "sha256": "eaed4c7f3d10a1ec975cb4e0e6fb230aa78919f619e64284adceec9adb0e567f",
+      "media_type": "text/csv",
+      "role": "dataset",
+      "source_release_path": "release/all-attempts.csv",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "data/all-attempts.jsonl",
+      "bytes": 718077,
+      "sha256": "0da584f9e98ea8c4a3791c0d5482d5ecd65ed59917daab8ba4e701be8b838860",
+      "media_type": "application/x-ndjson",
+      "role": "dataset",
+      "source_release_path": "release/all-attempts.jsonl",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "data/failures.csv",
+      "bytes": 50657,
+      "sha256": "d43f8d45885af48161c3c8fb4de8011c67f9ff85c5324e95e1f67942c54e8061",
+      "media_type": "text/csv",
+      "role": "dataset",
+      "source_release_path": "release/failures.csv",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "data/india-latency-cases.csv",
+      "bytes": 35122,
+      "sha256": "0046db3c972e085075bd109921d0439a6944ca5fe9e9532abc91670d118f0891",
+      "media_type": "text/csv",
+      "role": "dataset",
+      "source_release_path": "release/india-latency-cases.csv",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "data/india-matched-pairs.csv",
+      "bytes": 3708,
+      "sha256": "51c51e3cef225fef00117519a16742dbabb296721829a2e710a1b1896dddc0a5",
+      "media_type": "text/csv",
+      "role": "dataset",
+      "source_release_path": "release/india-matched-pairs.csv",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "data/model-fingerprints.csv",
+      "bytes": 1439,
+      "sha256": "5309cfd67f7d0fe930a02f57a32a4f0a5ab293fbaae1cfd0a4e4045c3618c527",
+      "media_type": "text/csv",
+      "role": "dataset",
+      "source_release_path": "release/model-fingerprints.csv",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "data/pilot-excluded-cases.csv",
+      "bytes": 20851,
+      "sha256": "ba4785785b734cc3c8ba0124617e54982f36bf3a8a13a5bd8299eb820f4a9721",
+      "media_type": "text/csv",
+      "role": "dataset",
+      "source_release_path": "release/pilot-excluded-cases.csv",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "data/primary-cases.csv",
+      "bytes": 278543,
+      "sha256": "c8c252b795cacd504eb360108e0ff71038eef195b87428cc753cb5c33d9df617",
+      "media_type": "text/csv",
+      "role": "dataset",
+      "source_release_path": "release/primary-cases.csv",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "data/release-case-inventory.csv",
+      "bytes": 61514,
+      "sha256": "642fa8c113da75d652e9ef961ef991aa898d03eaf361478bb18aae7402bf985c",
+      "media_type": "text/csv",
+      "role": "dataset",
+      "source_release_path": "release/release-case-inventory.csv",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "DATA-DICTIONARY.md",
+      "bytes": 4306,
+      "sha256": "20edca9b378257ef0c15ed39d52c8b72a30f5e81a106b0607ede03ebd2b42702",
+      "media_type": "text/markdown",
+      "role": "documentation",
+      "source_release_path": "release/DATA-DICTIONARY.md",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "LICENSE-DATA.txt",
+      "bytes": 103,
+      "sha256": "ef3418854f642053d285ea91f29121c23c95f1b672960da9ddb74f9c87a8fe28",
+      "media_type": "text/plain",
+      "role": "license",
+      "source_release_path": "release/LICENSE-DATA.txt",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/calibration.json",
+      "bytes": 93886,
+      "sha256": "0d8265a9f74bb39d9b1947e90e041dc330dcdb9b445d2d82a27da378f3bd88b2",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/calibration.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/methodology.json",
+      "bytes": 6541,
+      "sha256": "5b9edaf7898e95ad0d48a86d539c529d6be24ac816d625915fd80a52452987b0",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/methodology.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/pilot-evidence-summary.json",
+      "bytes": 591,
+      "sha256": "3284bf21cba423b516522ac1aca967334bfe185b4bb101d19c615498e954093c",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/pilot-evidence-summary.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/prices.json",
+      "bytes": 575,
+      "sha256": "3b36528039a263989c71369984415032e50f3e3bedfdb2778cf05ce53a770289",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/prices.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/protocol.json",
+      "bytes": 5017,
+      "sha256": "39c68bfc607157011012dc47d59524dcf3d5d7155071e6862c5038410d642c16",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/protocol.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/run-manifest.json",
+      "bytes": 1074,
+      "sha256": "e2c9b43ab2f70698f07a3147ff68f09befc94467925de6654ebddfc09c5592ad",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/run-manifest.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/summary.json",
+      "bytes": 82664,
+      "sha256": "c9b4b8576bec4827a36c0560f470174a600447695b5ac243006b16dd1f81d046",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/summary.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/upstream-checksums.sha256",
+      "bytes": 7410,
+      "sha256": "20436a28c19cf56839d7709c3e25d2a3d6e165d58738bfebf9df966c39b83b76",
+      "media_type": "text/plain",
+      "role": "provenance",
+      "source_release_path": "release/checksums.sha256",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/upstream-manifest.json",
+      "bytes": 13365,
+      "sha256": "5a20178b01c92c4cf6297750026da0677c2d6425364e649df34dd1692dfbcca4",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/manifest.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/upstream-qa-report.json",
+      "bytes": 1246,
+      "sha256": "753f351382093bea057599866be7665fc2246d0d59dd3075e55315359e3ae9b1",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/qa-report.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "provenance/validator-export-manifest.json",
+      "bytes": 905,
+      "sha256": "adc2f364e269ce63785cbca399a2329a25bba9c938e2b38f476b5ccaa3cb55f7",
+      "media_type": "application/json",
+      "role": "provenance",
+      "source_release_path": "release/validator-export-manifest.json",
+      "byte_preserved_from_source": true
+    },
+    {
+      "path": "CITATION.cff",
+      "bytes": 819,
+      "sha256": "39f0bfe8cca1c41e5fb2a4702cdf68289bfcfcc55135aca68ed21c7deec251f2",
+      "media_type": "text/yaml",
+      "role": "citation",
+      "source_release_path": null,
+      "byte_preserved_from_source": false
+    },
+    {
+      "path": "QA.json",
+      "bytes": 2760,
+      "sha256": "dfa046911fd7778e8288a47483bc5ac4179ee4601743e61706f586ea73ae39f3",
+      "media_type": "application/json",
+      "role": "bundle_qa",
+      "source_release_path": null,
+      "byte_preserved_from_source": false
+    },
+    {
+      "path": "README.md",
+      "bytes": 12157,
+      "sha256": "c16fcbda2a08a472c9df6b74e13dc985dfc118e303ea65761ae900fe20cae68c",
+      "media_type": "text/markdown",
+      "role": "dataset_card",
+      "source_release_path": null,
+      "byte_preserved_from_source": false
+    }
+  ]
+}

QA.json

@@ -0,0 +1,83 @@
+{
+  "schema_version": "1.0.0",
+  "status": "PASS_LOCAL_READY_FOR_UPLOAD",
+  "checked_on": "2026-08-07",
+  "external_upload_performed": false,
+  "source_release_version": "1.0.0",
+  "source_release_tag": "v1.0.0",
+  "source_release_commit": "06162a211ad9929646d53d60b97c31b758192d09",
+  "protocol_id": "deepseek-v4-long-context-retrieval-v1.1.0",
+  "doi": "10.5281/zenodo.21838863",
+  "doi_url": "https://doi.org/10.5281/zenodo.21838863",
+  "trusted_source_archive_sha256": "38337e1213e2cf7ba7de8f21703e11772eddb71326361a66f27f7a8d0c114dcb",
+  "expected_final_bundle_file_count": 26,
+  "manifest_scope_file_count": 25,
+  "checks": {
+    "source_files_selected": 22,
+    "source_files_byte_identical": 22,
+    "source_file_byte_mismatches": 0,
+    "readme_utf8_decodes_strictly": true,
+    "readme_has_yaml_frontmatter": true,
+    "yaml_top_level_fields": [
+      "license",
+      "language",
+      "pretty_name",
+      "size_categories",
+      "tags",
+      "configs"
+    ],
+    "yaml_known_top_level_fields_only": true,
+    "yaml_tab_characters": 0,
+    "yaml_configs_follow_hugging_face_data_files_shape": true,
+    "yaml_license_cc_by_4_0": true,
+    "yaml_language_en": true,
+    "config_count": 9,
+    "default_config": "primary",
+    "default_config_count": 1,
+    "config_paths_exist": 9,
+    "config_paths_missing": 0,
+    "config_row_counts_match": true,
+    "all_attempts_csv_rows": 344,
+    "all_attempts_jsonl_rows": 344,
+    "all_attempts_csv_jsonl_row_and_field_parity": true,
+    "all_attempts_csv_jsonl_mismatch_rows": 0,
+    "primary_rows": 288,
+    "pilot_excluded_rows": 20,
+    "india_latency_rows": 36,
+    "india_matched_pair_rows": 36,
+    "failure_rows": 154,
+    "model_fingerprint_rows": 8,
+    "case_inventory_rows": 344,
+    "doi_bound_in_readme": true,
+    "doi_url_bound_in_readme": true,
+    "doi_bound_in_citation_cff": true,
+    "doi_url_bound_in_citation_cff": true,
+    "doi_bound_in_manifest_metadata": true,
+    "doi_placeholder_occurrences": 0,
+    "unexpected_double_underscore_placeholders": 0,
+    "private_absolute_path_matches": 0,
+    "email_matches": 0,
+    "aws_arn_matches": 0,
+    "aws_access_key_matches": 0,
+    "bearer_token_matches": 0,
+    "private_chat_url_matches": 0,
+    "wordpress_admin_or_preview_url_matches": 0,
+    "arabic_script_matches": 0,
+    "manifest_excluded_from_its_own_entries": true
+  },
+  "config_rows": {
+    "primary": 288,
+    "all_attempts_csv": 344,
+    "all_attempts_jsonl": 344,
+    "pilot_excluded": 20,
+    "india_latency": 36,
+    "india_matched_pairs": 36,
+    "failures": 154,
+    "model_fingerprints": 8,
+    "case_inventory": 344
+  },
+  "maintenance_requirements": [
+    "Recompute MANIFEST.json and rerun the same QA checks after any byte change."
+  ],
+  "blockers": []
+}

README.md

@@ -1,3 +1,224 @@
 ---
 license: cc-by-4.0
+language:
+  - en
+pretty_name: DeepSeek 1M Context Benchmark v1.0.0
+size_categories:
+  - n<1K
+tags:
+  - deepseek
+  - long-context
+  - model-evaluation
+  - api-benchmark
+  - reproducibility
+  - synthetic-data
+configs:
+  - config_name: primary
+    default: true
+    data_files:
+      - split: test
+        path: data/primary-cases.csv
+  - config_name: all_attempts_csv
+    data_files:
+      - split: test
+        path: data/all-attempts.csv
+  - config_name: all_attempts_jsonl
+    data_files:
+      - split: test
+        path: data/all-attempts.jsonl
+  - config_name: pilot_excluded
+    data_files:
+      - split: test
+        path: data/pilot-excluded-cases.csv
+  - config_name: india_latency
+    data_files:
+      - split: test
+        path: data/india-latency-cases.csv
+  - config_name: india_matched_pairs
+    data_files:
+      - split: test
+        path: data/india-matched-pairs.csv
+  - config_name: failures
+    data_files:
+      - split: test
+        path: data/failures.csv
+  - config_name: model_fingerprints
+    data_files:
+      - split: test
+        path: data/model-fingerprints.csv
+  - config_name: case_inventory
+    data_files:
+      - split: test
+        path: data/release-case-inventory.csv
 ---
+
+# DeepSeek 1M Context Benchmark
+
+This dataset is the publication-safe measurement release for **DeepSeek 1M Context Benchmark: Retrieval Accuracy, Latency, and Cost**, version `v1.0.0`. It contains 344 sanitized terminal API records produced by the frozen protocol `deepseek-v4-long-context-retrieval-v1.1.0` during a bounded run from `2026-08-06T20:17:02.706Z` through `2026-08-07T00:07:44.737Z`.
+
+The study compared `deepseek-v4-flash` and `deepseek-v4-pro` on deterministic synthetic English retrieval and synthesis tasks. Provider-counted prompt tiers covered 32K, 128K, 512K, and approximately 950K tokens. The primary matrix varied model, prompt tier, objective task family, target position, and repeat while holding the request and grader contracts fixed.
+
+- **Versioned source:** [GitHub release v1.0.0](https://github.com/chatdeepai/deepseek-1m-context-benchmark/releases/tag/v1.0.0)
+- **Reproducible repository:** [chatdeepai/deepseek-1m-context-benchmark](https://github.com/chatdeepai/deepseek-1m-context-benchmark)
+- **Research article:** [DeepSeek 1M Context Benchmark](https://chat-deep.ai/research/deepseek-1m-context-benchmark/)
+- **Archival DOI:** [10.5281/zenodo.21838863](https://doi.org/10.5281/zenodo.21838863)
+
+## What is included
+
+The 344 terminal rows are divided into three analysis roles that must not be combined indiscriminately:
+
+| Role | Rows | Network vantage | Use in primary accuracy |
+|---|---:|---|---|
+| `primary_accuracy` | 288 | AWS `us-east-1` client vantage | Yes |
+| `pilot_excluded` | 20 | Two disclosed pilot vantages | No |
+| `india_latency_validation` | 36 | AWS `ap-south-1` client vantage | No |
+
+Only the 288 rows with `analysis_role=primary_accuracy` enter the published primary accuracy denominator. The pilot and India-vantage records are retained for auditability and separate bounded analyses.
+
+The release contains sanitized measurements and hashes. It does **not** publish raw prompts, raw request bodies, raw responses, raw SSE payloads, credentials, authorization material, cloud resource identifiers, account data, or private URLs.
+
+## Configurations
+
+Every configuration uses the Hugging Face `test` split because these are measured benchmark outcomes, not training examples.
+
+| Config | File | Rows | Columns | Purpose |
+|---|---|---:|---:|---|
+| `primary` | `data/primary-cases.csv` | 288 | 51 | Default config and only primary accuracy denominator. |
+| `all_attempts_csv` | `data/all-attempts.csv` | 344 | 51 | All sanitized terminal rows in CSV format. |
+| `all_attempts_jsonl` | `data/all-attempts.jsonl` | 344 | 51 | The same 344 terminal rows in JSON Lines format. |
+| `pilot_excluded` | `data/pilot-excluded-cases.csv` | 20 | 51 | Paid pilot calls, excluded from primary accuracy. |
+| `india_latency` | `data/india-latency-cases.csv` | 36 | 51 | Bounded India client-network-vantage validation rows. |
+| `india_matched_pairs` | `data/india-matched-pairs.csv` | 36 | 15 | Matched U.S.-versus-India timing comparisons. |
+| `failures` | `data/failures.csv` | 154 | 15 | Non-exact terminal states across all analysis roles. |
+| `model_fingerprints` | `data/model-fingerprints.csv` | 8 | 8 | Requested/returned model and fingerprint group counts. |
+| `case_inventory` | `data/release-case-inventory.csv` | 344 | 11 | Presentation-sorted inventory of released cases. |
+
+`all_attempts_csv` and `all_attempts_jsonl` are alternate serializations of the same records. Do not concatenate them. Empty fields mean that the provider or adapter did not supply a value; no missing values were silently imputed. See [DATA-DICTIONARY.md](DATA-DICTIONARY.md) for the 51-field case schema.
+
+## Loading the dataset
+
+After publication, replace the example repository identifier with the final Hugging Face dataset ID:
+
+```python
+from datasets import load_dataset
+
+dataset_id = "replace-with-the-final-hugging-face-dataset-id"
+
+# The 288-case primary benchmark matrix.
+primary = load_dataset(dataset_id, "primary", split="test")
+
+# All 344 terminal rows in CSV form.
+all_attempts = load_dataset(dataset_id, "all_attempts_csv", split="test")
+
+# The bounded India client-network-vantage subset.
+india = load_dataset(dataset_id, "india_latency", split="test")
+```
+
+For a local checkout, the underlying files can also be loaded directly with the `csv` or `json` builders.
+
+## Benchmark design
+
+The frozen 288-case primary matrix is:
+
+- 2 exact API model IDs: `deepseek-v4-flash` and `deepseek-v4-pro`;
+- 4 provider-counted prompt tiers: 32K, 128K, 512K, and approximately 950K;
+- 4 objective task families: single-record retrieval, two-record join, latest-version conflict resolution, and scattered event ordering;
+- 3 target-position labels: beginning, middle, and end;
+- 3 deterministic fixture repeats.
+
+The benchmark used streaming Chat Completions, non-thinking mode, temperature `0`, JSON Output, a 256-token study output cap, and a strict family-specific grader. Exact match required the expected key set, types, and values with no surrounding prose. The 256-token value is this study's generation cap, not a claim about the provider's maximum supported output.
+
+The 20 pilot calls are excluded from all primary denominators. The 36 India-vantage calls form a separate matched client-network-vantage check at the 32K and approximately 950K edges; they do not represent Indian users or identify provider hosting locations.
+
+## Selected versioned results
+
+These figures describe release `v1.0.0` only:
+
+- Primary strict exact match: **152/288 (52.78%)**.
+- V4 Flash strict exact match: **71/144 (49.31%)**.
+- V4 Pro strict exact match: **81/144 (56.25%)**.
+- Primary valid JSON and exact key-set rates: **288/288 (100%)** each.
+- Primary end-to-end stream time: **12,748 ms p50** and **162,295 ms p95** across the complete primary matrix.
+- Dated primary cache-miss cost upper bound: **USD 33.589136**, using the official price snapshot frozen on 2026-08-06 and complete returned usage for 288/288 rows.
+
+These results are task-specific. They do not establish a universal model winner, and the dated cost figures are not current price quotes.
+
+## Provenance and integrity
+
+The public release was built through an explicit allowlist from a validator export that reconciled all 344 planned terminal rows. The trusted publication-safe source archive has SHA-256:
+
+```text
+38337e1213e2cf7ba7de8f21703e11772eddb71326361a66f27f7a8d0c114dcb
+```
+
+The following files preserve the versioned methodology and release evidence:
+
+- [`provenance/protocol.json`](provenance/protocol.json): frozen experimental contract.
+- [`provenance/calibration.json`](provenance/calibration.json): deterministic prompt-tier calibration.
+- [`provenance/methodology.json`](provenance/methodology.json): publication methodology and analysis boundaries.
+- [`provenance/summary.json`](provenance/summary.json): canonical versioned aggregate metrics.
+- [`provenance/prices.json`](provenance/prices.json): dated price snapshot used for reported estimates.
+- [`provenance/pilot-evidence-summary.json`](provenance/pilot-evidence-summary.json): public-safe pilot and preflight disclosure.
+- [`provenance/run-manifest.json`](provenance/run-manifest.json) and [`provenance/validator-export-manifest.json`](provenance/validator-export-manifest.json): frozen run and export evidence.
+- [`provenance/upstream-manifest.json`](provenance/upstream-manifest.json), [`provenance/upstream-qa-report.json`](provenance/upstream-qa-report.json), and [`provenance/upstream-checksums.sha256`](provenance/upstream-checksums.sha256): integrity records for the complete GitHub release. The upstream checksum list also names charts and tables that are intentionally not duplicated in this compact Hugging Face bundle.
+- [`CITATION.cff`](CITATION.cff): machine-readable citation metadata bound to the versioned Zenodo DOI.
+- [`MANIFEST.json`](MANIFEST.json) and [`QA.json`](QA.json): integrity and safety checks for this Hugging Face upload bundle itself.
+
+The full fixture generator, grader, publication code, charts, and release tables remain in the versioned GitHub repository.
+
+## Intended uses
+
+This dataset is suitable for:
+
+- auditing the published primary accuracy calculation;
+- reanalyzing accuracy by model, prompt tier, task family, and target position;
+- inspecting transport, streaming, token, cache, and dated cost telemetry;
+- reproducing published tables or creating new visualizations;
+- studying matched outcomes and bounded client-network-vantage latency differences;
+- teaching reproducible benchmark reporting, denominator separation, and release provenance.
+
+## Out-of-scope uses
+
+This dataset should not be treated as:
+
+- a training or fine-tuning corpus;
+- a general intelligence, coding, factuality, safety, multilingual, agent, or web-search benchmark;
+- a representative sample of real user prompts;
+- evidence of provider server location;
+- a recurring reliability or availability monitor;
+- a current pricing source;
+- a source of raw prompts or raw model outputs.
+
+## Limitations and potential sources of bias
+
+- All benchmark records are deterministic synthetic English data.
+- The strict JSON grader measures narrow exact retrieval and synthesis behavior, not open-ended response quality.
+- The run covers one bounded execution window and two exact model IDs as routed during that window.
+- Target-position semantics differ by task family, so aggregate position comparisons require the detailed methodology.
+- Latency includes the disclosed client network vantage and workload; it should not be generalized to every user, region, or date.
+- The India subset is intentionally small and limited to matched single-record cases at two prompt tiers.
+- Provider pricing, model aliases, routing, and fingerprints can change after this release.
+- Hashes of withheld raw artifacts support provenance but do not make the raw content public.
+
+## Licensing and attribution
+
+The dataset is licensed under [Creative Commons Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/). See [LICENSE-DATA.txt](LICENSE-DATA.txt). Attribute **Chat Deep AI**, identify version `v1.0.0`, link the versioned release, and retain material methodology and limitation notices when redistributing or adapting the data.
+
+Provider names and trademarks belong to their respective owners. This is an independent benchmark and is not an official DeepSeek dataset.
+
+## Citation
+
+Use the versioned archival DOI `10.5281/zenodo.21838863`:
+
+```bibtex
+@dataset{chat_deep_ai_2026_deepseek_1m_context,
+  author    = {Chat Deep AI},
+  title     = {DeepSeek 1M Context Benchmark: Retrieval Accuracy, Latency, and Cost},
+  year      = {2026},
+  version   = {1.0.0},
+  doi       = {10.5281/zenodo.21838863},
+  url       = {https://github.com/chatdeepai/deepseek-1m-context-benchmark/releases/tag/v1.0.0}
+}
+```
+
+The archival record resolves at [https://doi.org/10.5281/zenodo.21838863](https://doi.org/10.5281/zenodo.21838863).