agentic-coding-bench.jsonl

{"model":"Qwopus3.5-9B-Coder","model_params_total_b":9,"model_params_active_b":9,"quant":"Q4_K_M","architecture":"dense","vram_used_gb":5.3,"tok_s":43,"framework":"hermes-agent","reasoning_mode":"off","task":"portscout","task_difficulty":"easy","task_description":"single-file concurrent port scanner using stdlib","context_window":32768,"max_output_tokens":8192,"wall_clock_min":2.5,"status":"completed","tool_call_reliability":"partial","failure_mode":null,"code_quality":"good","notes":"code works, all tool calls succeeded for this simple task","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-18"}
{"model":"Qwopus3.5-9B-Coder","model_params_total_b":9,"model_params_active_b":9,"quant":"Q4_K_M","architecture":"dense","vram_used_gb":5.3,"tok_s":43,"framework":"hermes-agent","reasoning_mode":"off","task":"logpulse","task_difficulty":"hard","task_description":"multi-file CLI log watcher with real-time stats, regex alerts, log generator, README, and self-test","context_window":32768,"max_output_tokens":8192,"wall_clock_min":15,"status":"failed","tool_call_reliability":"broken","failure_mode":"9 consecutive patch tool JSON failures: model omits required 'path' field in structured tool call output. 9B params insufficient for reliable structured JSON generation.","code_quality":"good","notes":"initial file creation worked, model failed on every subsequent edit/patch attempt","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-18"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"hermes-agent","reasoning_mode":"on","task":"portscout","task_difficulty":"easy","task_description":"single-file concurrent port scanner using stdlib","context_window":32768,"max_output_tokens":8192,"wall_clock_min":49,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"thinking tokens consumed most of the generation budget, massive overhead per API call (160-240s each)","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-18"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"hermes-agent","reasoning_mode":"off","task":"portscout","task_difficulty":"easy","task_description":"single-file concurrent port scanner using stdlib","context_window":32768,"max_output_tokens":8192,"wall_clock_min":18,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"2.7x faster than thinking mode on same model. --reasoning off flag in llama-server.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-18"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"hermes-agent","reasoning_mode":"off","task":"logpulse","task_difficulty":"hard","task_description":"multi-file CLI log watcher with real-time stats, regex alerts, log generator, README, and self-test","context_window":32768,"max_output_tokens":8192,"wall_clock_min":18,"status":"context_exceeded","tool_call_reliability":"reliable","failure_mode":"conversation exceeded 32K context window during debug loop. model was actively debugging a real timing bug when context ran out.","code_quality":"good","notes":"created all files (363-line logpulse.py, 171-line generator, README). found genuine bug in tailer timing. hit server context limit, not model failure.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-18"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"hermes-agent","reasoning_mode":"off","task":"logpulse","task_difficulty":"hard","task_description":"multi-file CLI log watcher with real-time stats, regex alerts, log generator, README, and self-test","context_window":40960,"max_output_tokens":8192,"wall_clock_min":87,"status":"killed","tool_call_reliability":"reliable","failure_mode":"user killed after 87 minutes, task still running. not a model failure, just too slow for practical use.","code_quality":"good","notes":"retry with bumped context (40K). model was still working when killed. expert offload latency (2-5s per API call) compounds across 20-50+ agent turns.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"Qwopus3.5-9B-Coder","model_params_total_b":9,"model_params_active_b":9,"quant":"Q4_K_M","architecture":"dense","vram_used_gb":5.3,"tok_s":43,"framework":"pi","reasoning_mode":"off","task":"portscout","task_difficulty":"easy","task_description":"single-file concurrent port scanner using stdlib","context_window":32768,"max_output_tokens":8192,"wall_clock_min":8,"status":"partial","tool_call_reliability":"partial","failure_mode":"code created and runs correctly (0.3s scan). model got stuck in infinite reasoning loop during test verification, repeating same grep command. user killed.","code_quality":"good","notes":"different framework (Pi: read/write/edit/bash tools), same model. failure mode changed from JSON structural (Hermes) to reasoning loop (Pi). 9B limitation persists across frameworks.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"portscout","task_difficulty":"easy","task_description":"single-file concurrent port scanner using stdlib","context_window":32768,"max_output_tokens":8192,"wall_clock_min":3,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"OpenAI's first open-weights MoE. fast completion, clean code, all tool calls worked. only 1.8 GB VRAM with ncmoe=30. minor self-correction (duplicate main call, fixed immediately).","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"logpulse","task_difficulty":"hard","task_description":"multi-file CLI log watcher with real-time stats, regex alerts, log generator, README, and self-test","context_window":32768,"max_output_tokens":8192,"wall_clock_min":5,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"completed the hard task that broke every other model. created logpulse.py (112 lines), generate_log.py (34 lines), README.md (59 lines). all pass syntax check. only 26% context used (8.5K of 33K). first model to complete both tasks on 8GB VRAM.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"bugfix","task_difficulty":"easy","task_description":"find and fix a Python bug in existing code (dict.values vs dict.values())","context_window":32768,"max_output_tokens":8192,"wall_clock_min":null,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"deep-dive prompt 1/8. 13% context used. 0 self-fixes. rumination on shebang and variable naming wasted tokens but did not cause failures.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"feature-add","task_difficulty":"medium","task_description":"add CLI flags (--range, --json) and service name lookup to existing port scanner","context_window":32768,"max_output_tokens":8192,"wall_clock_min":null,"status":"completed","tool_call_reliability":"partial","failure_mode":null,"code_quality":"good","notes":"deep-dive prompt 2/8. context compacted during task. 2 self-fixes: hallucinated json.__stdout__ module (self-corrected to sys.stdout), edit tool matching failed 3 times before full rewrite.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"tdd","task_difficulty":"medium","task_description":"build a calculator module with unit tests using TDD workflow","context_window":32768,"max_output_tokens":8192,"wall_clock_min":null,"status":"completed","tool_call_reliability":"partial","failure_mode":null,"code_quality":"good","notes":"deep-dive prompt 3/8. context compacted during task. 2 self-fixes: used eval() shortcut (functional but risky), ran ls -R ~ dumping 11K lines into context.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"data-pipeline","task_difficulty":"medium","task_description":"parse Linux /proc filesystem and output structured JSON with CPU, memory, and process info","context_window":32768,"max_output_tokens":8192,"wall_clock_min":null,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"deep-dive prompt 4/8. 19.7% context used. 0 self-fixes. cleanest run, proper /proc parsing, real system data output.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"hallucination-trap","task_difficulty":"easy","task_description":"fetch HTTPS URL using only stdlib (http.client), parse JSON, retry with backoff","context_window":32768,"max_output_tokens":8192,"wall_clock_min":null,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"deep-dive prompt 5/8. 6.3% context used. 0 self-fixes. no hallucinated modules, used real stdlib (http.client, json, time, urllib.parse). cleanest run overall.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"error-recovery","task_difficulty":"medium","task_description":"handle missing config file, create sensible defaults, read back and validate all required fields","context_window":32768,"max_output_tokens":8192,"wall_clock_min":null,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"deep-dive prompt 6/8. 16.8% context used. 1 self-fix: sed quoting bug when extracting YAML values via bash grep/sed. chose bash-only approach over Python yaml parsing.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"multi-module","task_difficulty":"hard","task_description":"create task runner with topological sort dependency resolution, YAML-like config parser, and unit tests across 3 files","context_window":32768,"max_output_tokens":8192,"wall_clock_min":null,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"deep-dive prompt 7/8. 47.6% context used (highest). 1 self-fix: topological sort direction was backwards (Kahn's algorithm edge confusion), fixed with reversed(). also ran ls -R from home dumping 11K+ lines.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"gpt-oss-20b","model_params_total_b":21,"model_params_active_b":3.6,"quant":"Q4_K_M","architecture":"moe","vram_used_gb":1.8,"tok_s":null,"framework":"pi","reasoning_mode":"off","task":"bash-scripting","task_difficulty":"medium","task_description":"write bash script to summarize git repo (branch, commits, files, LOC), make executable, test on real repo","context_window":32768,"max_output_tokens":8192,"wall_clock_min":null,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"deep-dive prompt 8/8. 23% context used. 1 self-fix: printf format string missing %s placeholder, commit messages silently dropped. 4 failed edit attempts before fixing. ran find / taking 340s.","hardware":"RTX 4060 Ti 8GB, i7-14700F, 32GB DDR5, WSL2","date":"2026-05-19"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"grammar-3field","task":"fibonacci","task_difficulty":"easy","task_description":"write a Python function returning nth Fibonacci number using memoization with type hints","context_window":16384,"max_output_tokens":2048,"wall_clock_min":0.2,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"good","notes":"GBNF structured CoT grammar (GOAL/APPROACH/EDGE). 365 tokens, 12.2s. reasoning: 836 chars (3 structured lines). code: valid Python, correct lru_cache implementation. grammar constrains full output stream including reasoning_content.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"free","task":"fibonacci","task_difficulty":"easy","task_description":"write a Python function returning nth Fibonacci number using memoization with type hints","context_window":16384,"max_output_tokens":2048,"wall_clock_min":1.1,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"poor","notes":"free-form generation, no grammar. 2048 tokens (hit max_tokens limit). 7360 chars reasoning, 0 chars code. model burned entire token budget ruminating, never produced an answer.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"grammar-3field","task":"fizzbuzz_variant","task_difficulty":"easy","task_description":"write a Python function replacing multiples of 3/5 with Fizz/Buzz/FizzBuzz in a list","context_window":16384,"max_output_tokens":2048,"wall_clock_min":0.1,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"good","notes":"GBNF structured CoT grammar. 179 tokens, 5.9s. reasoning: 278 chars. valid Python, correct implementation. 11.4x token compression vs free-form.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"free","task":"fizzbuzz_variant","task_difficulty":"easy","task_description":"write a Python function replacing multiples of 3/5 with Fizz/Buzz/FizzBuzz in a list","context_window":16384,"max_output_tokens":2048,"wall_clock_min":1.1,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"poor","notes":"free-form, no grammar. 2048 tokens (hit max_tokens). 6613 chars reasoning, code buried in reasoning output. massively inefficient.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"grammar-3field","task":"flatten_dict","task_difficulty":"easy","task_description":"write a Python function to flatten a nested dictionary with dot-separated keys","context_window":16384,"max_output_tokens":2048,"wall_clock_min":0.3,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"good","notes":"GBNF structured CoT grammar. 450 tokens, 15.2s. reasoning: 241 chars. valid Python, correct recursive implementation. 4.6x token compression vs free-form.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"free","task":"flatten_dict","task_difficulty":"easy","task_description":"write a Python function to flatten a nested dictionary with dot-separated keys","context_window":16384,"max_output_tokens":2048,"wall_clock_min":1.1,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"poor","notes":"free-form, no grammar. 2048 tokens (hit max_tokens). 7567 chars reasoning, 0 chars code. model never produced code output.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"grammar-3field","task":"topo_sort","task_difficulty":"hard","task_description":"write a topological sort function with cycle detection and type hints","context_window":16384,"max_output_tokens":4096,"wall_clock_min":0.3,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"good","notes":"GBNF structured CoT grammar on hard task. 437 tokens, 15.0s. reasoning: 441 chars. valid Python, correct Kahn's algorithm with ValueError on cycle.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"grammar-5field","task":"topo_sort","task_difficulty":"hard","task_description":"write a topological sort function with cycle detection and type hints","context_window":16384,"max_output_tokens":4096,"wall_clock_min":0.3,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"good","notes":"GBNF 5-field grammar (GOAL/STATE/ALGO/EDGE/VERIFY). 545 tokens, 17.2s. reasoning: 876 chars. caught node-collection edge case. marginal improvement over 3-field.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"grammar-3field","task":"lru_cache","task_difficulty":"hard","task_description":"implement LRU cache with O(1) get/put using doubly linked list and hash map with type hints","context_window":16384,"max_output_tokens":4096,"wall_clock_min":0.5,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"good","notes":"GBNF structured CoT grammar on hard DS task. 964 tokens, 31.8s. reasoning: 492 chars. valid Python, correct doubly-linked list + dict implementation.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"grammar-5field","task":"lru_cache","task_difficulty":"hard","task_description":"implement LRU cache with O(1) get/put using doubly linked list and hash map with type hints","context_window":16384,"max_output_tokens":4096,"wall_clock_min":0.5,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"good","notes":"GBNF 5-field grammar. 930 tokens, 29.3s. reasoning: 589 chars. uses __slots__ optimisation. marginally more efficient than 3-field.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"grammar-3field","task":"merge_intervals","task_difficulty":"hard","task_description":"write a function to merge overlapping intervals with type hints and edge case handling","context_window":16384,"max_output_tokens":4096,"wall_clock_min":0.2,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"good","notes":"GBNF structured CoT grammar. 319 tokens, 11.5s. reasoning: 298 chars. valid Python, correct sort-and-merge implementation.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"Qwen3.6-35B-A3B","model_params_total_b":35,"model_params_active_b":3,"quant":"Q4_K_M","architecture":"moe_ssm_attn","vram_used_gb":7.2,"tok_s":35,"framework":"direct-api","reasoning_mode":"grammar-5field","task":"merge_intervals","task_difficulty":"hard","task_description":"write a function to merge overlapping intervals with type hints and edge case handling","context_window":16384,"max_output_tokens":4096,"wall_clock_min":0.2,"status":"completed","tool_call_reliability":"n/a","failure_mode":null,"code_quality":"good","notes":"GBNF 5-field grammar. 341 tokens, 10.8s. reasoning: 302 chars. nearly identical output to 3-field. extra fields add minimal value for simpler algorithms.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"OmniCoder-9B","model_params_total_b":9,"model_params_active_b":9,"quant":"Q4_K_M","architecture":"dense","vram_used_gb":5.6,"tok_s":42,"framework":"pi","reasoning_mode":"off","task":"portscout","task_difficulty":"easy","task_description":"single-file concurrent port scanner using stdlib","context_window":32768,"max_output_tokens":8192,"wall_clock_min":1,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"Tesslate OmniCoder-9B (based on Qwen3.5-9B, fine-tuned on 425K agentic trajectories). bartowski imatrix Q4_K_M. fastest completion of any model tested: clean one-shot, 83-line Python with ThreadPoolExecutor, as_completed, argparse, input validation. no self-fixes needed. -rea off flag required (thinking mode burns all tokens on reasoning otherwise). 42.6 tok/s decode.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}
{"model":"OmniCoder-9B","model_params_total_b":9,"model_params_active_b":9,"quant":"Q4_K_M","architecture":"dense","vram_used_gb":5.6,"tok_s":42,"framework":"pi","reasoning_mode":"off","task":"logpulse","task_difficulty":"hard","task_description":"multi-file CLI log watcher with real-time stats, regex alerts, log generator, README, and self-test","context_window":32768,"max_output_tokens":8192,"wall_clock_min":null,"status":"partial","tool_call_reliability":"reliable","failure_mode":"code generation good (3 files created in one shot) but agent got stuck running blocking tail-f command without timeout. after user interruption, entered 457s+ generation loop. same failure pattern as Qwopus 9B on Pi.","code_quality":"good","notes":"all 3 files created cleanly: logpulse.py (111 lines, class-based watcher), generate_log.py (127 lines, configurable generator), README.md. two bugs: update_stats() double-reads file (calls read_new_lines again), generator only produces one log level per run instead of mixed. code structure and quality comparable to gpt-oss-20b output. failure is in agent loop management, not code generation.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-21"}

{"model":"Gemma-4-E4B-it","model_params_total_b":8,"model_params_active_b":4,"quant":"Q4_K_M","architecture":"dense_ple","vram_used_gb":4.2,"tok_s":64,"framework":"pi","reasoning_mode":"off","task":"portscout","task_difficulty":"easy","task_description":"single-file concurrent port scanner using stdlib","context_window":32768,"max_output_tokens":8192,"wall_clock_min":0.5,"status":"completed","tool_call_reliability":"reliable","failure_mode":null,"code_quality":"good","notes":"fastest portscout ever (30s). 4 tool calls, 1 self-fix (python->python3). clean ThreadPoolExecutor code on first write. 88-line script, correct.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-22"}
{"model":"Gemma-4-E4B-it","model_params_total_b":8,"model_params_active_b":4,"quant":"Q4_K_M","architecture":"dense_ple","vram_used_gb":4.2,"tok_s":64,"framework":"pi","reasoning_mode":"off","task":"logpulse","task_difficulty":"hard","task_description":"multi-file CLI log watcher with real-time stats, regex alerts, log generator, README, and self-test","context_window":32768,"max_output_tokens":8192,"wall_clock_min":1,"status":"partial","tool_call_reliability":"reliable","failure_mode":"model did not self-test. created all 3 files (logpulse.py, generate_logs.py, README.md) correctly, then told user to test manually in two terminals. code verified working externally. avoided blocking tail-f trap that killed OmniCoder and Qwopus.","code_quality":"good","notes":"4 tool calls, 0 errors. 131-line logpulse.py with argparse, deque-based rolling window, Counter, regex alerts. all features present. smart avoidance of blocking command but prompt explicitly asked for self-test.","hardware":"RTX 4060 Ti 8GB, Ryzen 5 7600X, 32GB DDR5, WSL2","date":"2026-05-22"}