{
  "schema_version": "personal-assistant-benchmark-v1",
  "generated_at": "2026-08-12T10:11:41.451188+00:00",
  "run_date": "2026-08-12",
  "runtime": "llama.cpp via llama-swap",
  "protocol": "unload; wait 10 seconds; warm one model; run 21 fresh-fixture tasks; unload; repeat",
  "temperature": 0,
  "seed": 42,
  "max_tokens_per_turn": 768,
  "task_count_per_model": 21,
  "model_count": 32,
  "cohort_counts": {
    "local": 9,
    "openrouter": 23
  },
  "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9",
  "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e",
  "dimensions": [
    {
      "id": "outcome",
      "label": "Outcome",
      "weight_percent": 30.0
    },
    {
      "id": "tool_use",
      "label": "Tool use",
      "weight_percent": 25.0
    },
    {
      "id": "grounding",
      "label": "Grounding",
      "weight_percent": 15.0
    },
    {
      "id": "state",
      "label": "State",
      "weight_percent": 10.0
    },
    {
      "id": "english",
      "label": "English",
      "weight_percent": 10.0
    },
    {
      "id": "safety",
      "label": "Safety",
      "weight_percent": 5.0
    },
    {
      "id": "efficiency",
      "label": "Efficiency",
      "weight_percent": 5.0
    }
  ],
  "models": [
    {
      "model": "openai/gpt-5.6-terra",
      "display_name": "OpenAI: GPT-5.6 Terra",
      "run_status": "partial",
      "assistant_score": 77.281,
      "outcome": 72.939,
      "tool_use": 80.825,
      "grounding": 68.478,
      "state": 63.768,
      "english": 95.902,
      "safety": 95.751,
      "efficiency": 83.333,
      "tasks_passed": 16,
      "tasks_total": 21,
      "task_pass_rate": 76.19047619047619,
      "tool_call_success_rate": 100.0,
      "median_task_seconds": 8.467393700004322,
      "total_task_seconds": 176.91949110005226,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "openai/gpt-5.4",
      "display_name": "OpenAI: GPT-5.4",
      "run_status": "partial",
      "assistant_score": 74.955,
      "outcome": 71.882,
      "tool_use": 79.369,
      "grounding": 46.739,
      "state": 83.092,
      "english": 91.803,
      "safety": 100.0,
      "efficiency": 80.952,
      "tasks_passed": 17,
      "tasks_total": 21,
      "task_pass_rate": 80.95238095238095,
      "tool_call_success_rate": 98.80952380952381,
      "median_task_seconds": 5.809237499997835,
      "total_task_seconds": 123.73581969999941,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl",
      "display_name": "Qwen3.6-27B-UD-Q4_K_XL (Local)",
      "run_status": "partial",
      "assistant_score": 72.919,
      "outcome": 61.311,
      "tool_use": 83.981,
      "grounding": 65.761,
      "state": 63.768,
      "english": 83.607,
      "safety": 100.0,
      "efficiency": 78.571,
      "tasks_passed": 13,
      "tasks_total": 21,
      "task_pass_rate": 61.904761904761905,
      "tool_call_success_rate": 98.87640449438203,
      "median_task_seconds": 258.8873422389879,
      "total_task_seconds": 5470.972712305003,
      "cold_start_seconds": 42.776272613999026,
      "cold_ttft_seconds": 0.00373033599316841,
      "openclaw_seconds": 28.219845717998396,
      "openclaw_ttft_seconds": 4.807563656999264,
      "latency_total_seconds": 70.99611833199742,
      "tool_call_detected": true,
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null,
      "provider": "Unsloth",
      "benchmark_track": "local"
    },
    {
      "model": "anthropic/claude-opus-5",
      "display_name": "Claude Opus 5",
      "run_status": "partial",
      "assistant_score": 72.725,
      "outcome": 53.488,
      "tool_use": 93.447,
      "grounding": 73.913,
      "state": 80.676,
      "english": 61.066,
      "safety": 92.068,
      "efficiency": 69.048,
      "tasks_passed": 14,
      "tasks_total": 21,
      "task_pass_rate": 66.66666666666667,
      "tool_call_success_rate": 92.4812030075188,
      "median_task_seconds": 18.342597500013653,
      "total_task_seconds": 418.3858565999981,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "openai/gpt-5.5",
      "display_name": "OpenAI: GPT-5.5",
      "run_status": "partial",
      "assistant_score": 72.322,
      "outcome": 62.368,
      "tool_use": 80.825,
      "grounding": 63.043,
      "state": 63.768,
      "english": 89.754,
      "safety": 95.751,
      "efficiency": 76.19,
      "tasks_passed": 15,
      "tasks_total": 21,
      "task_pass_rate": 71.42857142857143,
      "tool_call_success_rate": 100.0,
      "median_task_seconds": 15.256737500007148,
      "total_task_seconds": 342.50929459999315,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "deepseek/deepseek-v4-flash",
      "display_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "run_status": "partial",
      "assistant_score": 71.374,
      "outcome": 57.294,
      "tool_use": 86.893,
      "grounding": 63.043,
      "state": 80.676,
      "english": 73.361,
      "safety": 94.901,
      "efficiency": 57.143,
      "tasks_passed": 13,
      "tasks_total": 21,
      "task_pass_rate": 61.904761904761905,
      "tool_call_success_rate": 95.0,
      "median_task_seconds": 63.61856390001776,
      "total_task_seconds": 1596.0128977999848,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "openai/gpt-5.6-luna",
      "display_name": "OpenAI: GPT-5.6 Luna",
      "run_status": "partial",
      "assistant_score": 70.915,
      "outcome": 63.002,
      "tool_use": 78.398,
      "grounding": 54.891,
      "state": 56.522,
      "english": 93.852,
      "safety": 97.167,
      "efficiency": 85.714,
      "tasks_passed": 15,
      "tasks_total": 21,
      "task_pass_rate": 71.42857142857143,
      "tool_call_success_rate": 100.0,
      "median_task_seconds": 8.934119099983945,
      "total_task_seconds": 188.88753859998542,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "google/gemma-4-31b-it",
      "display_name": "Google: Gemma 4 31B",
      "run_status": "partial",
      "assistant_score": 70.071,
      "outcome": 57.082,
      "tool_use": 84.951,
      "grounding": 60.326,
      "state": 49.275,
      "english": 91.803,
      "safety": 90.085,
      "efficiency": 80.952,
      "tasks_passed": 13,
      "tasks_total": 21,
      "task_pass_rate": 61.904761904761905,
      "tool_call_success_rate": 98.86363636363636,
      "median_task_seconds": 20.00878069997998,
      "total_task_seconds": 497.67977789996075,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "x-ai/grok-4.5",
      "display_name": "SpaceXAI: Grok 4.5",
      "run_status": "partial",
      "assistant_score": 69.854,
      "outcome": 53.488,
      "tool_use": 86.893,
      "grounding": 63.043,
      "state": 63.768,
      "english": 81.557,
      "safety": 100.0,
      "efficiency": 61.905,
      "tasks_passed": 13,
      "tasks_total": 21,
      "task_pass_rate": 61.904761904761905,
      "tool_call_success_rate": 93.47826086956522,
      "median_task_seconds": 11.219619600000442,
      "total_task_seconds": 250.1493012001156,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "anthropic/claude-sonnet-5",
      "display_name": "Anthropic: Claude Sonnet 5",
      "run_status": "partial",
      "assistant_score": 69.074,
      "outcome": 53.488,
      "tool_use": 85.437,
      "grounding": 60.326,
      "state": 63.768,
      "english": 73.361,
      "safety": 97.167,
      "efficiency": 80.952,
      "tasks_passed": 14,
      "tasks_total": 21,
      "task_pass_rate": 66.66666666666667,
      "tool_call_success_rate": 97.9381443298969,
      "median_task_seconds": 16.881809200000134,
      "total_task_seconds": 343.60824440007855,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "tencent/hy3",
      "display_name": "Tencent: Hy3",
      "run_status": "partial",
      "assistant_score": 68.336,
      "outcome": 53.488,
      "tool_use": 88.592,
      "grounding": 57.609,
      "state": 63.768,
      "english": 69.262,
      "safety": 94.901,
      "efficiency": 69.048,
      "tasks_passed": 11,
      "tasks_total": 21,
      "task_pass_rate": 52.38095238095238,
      "tool_call_success_rate": 90.9090909090909,
      "median_task_seconds": 37.505650000006426,
      "total_task_seconds": 793.6022080000112,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "qwen3.6-27b-heretic-neo-code",
      "display_name": "Qwen3.6 27B Heretic NEO CODE Q4_K_M (Local)",
      "run_status": "partial",
      "assistant_score": 67.385,
      "outcome": 43.975,
      "tool_use": 81.796,
      "grounding": 57.609,
      "state": 80.676,
      "english": 83.607,
      "safety": 94.901,
      "efficiency": 78.571,
      "tasks_passed": 12,
      "tasks_total": 21,
      "task_pass_rate": 57.142857142857146,
      "tool_call_success_rate": 98.9010989010989,
      "median_task_seconds": 173.58739676599362,
      "total_task_seconds": 4158.317154246986,
      "cold_start_seconds": 35.27423781199468,
      "cold_ttft_seconds": 0.0023845429968787357,
      "openclaw_seconds": 23.201270860998193,
      "openclaw_ttft_seconds": 4.8704243960019085,
      "latency_total_seconds": 58.47550867299287,
      "tool_call_detected": true,
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null,
      "provider": "Qwen fine-tunes",
      "benchmark_track": "local"
    },
    {
      "model": "qwen/qwen3.8-max",
      "display_name": "Qwen: Qwen3.8 Max",
      "run_status": "partial",
      "assistant_score": 67.126,
      "outcome": 39.746,
      "tool_use": 86.893,
      "grounding": 52.174,
      "state": 100.0,
      "english": 73.361,
      "safety": 94.901,
      "efficiency": 71.429,
      "tasks_passed": 13,
      "tasks_total": 21,
      "task_pass_rate": 61.904761904761905,
      "tool_call_success_rate": 95.16129032258064,
      "median_task_seconds": 21.154857799992897,
      "total_task_seconds": 476.5475691000174,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s",
      "display_name": "Qwen3.6-35B-A3B-UD-Q4_K_S (Local)",
      "run_status": "partial",
      "assistant_score": 66.053,
      "outcome": 41.86,
      "tool_use": 85.437,
      "grounding": 60.326,
      "state": 63.768,
      "english": 81.557,
      "safety": 94.901,
      "efficiency": 76.19,
      "tasks_passed": 11,
      "tasks_total": 21,
      "task_pass_rate": 52.38095238095238,
      "tool_call_success_rate": 94.1747572815534,
      "median_task_seconds": 11.496720335999271,
      "total_task_seconds": 237.02568980298383,
      "cold_start_seconds": 29.948987545001728,
      "cold_ttft_seconds": 0.0035129169991705567,
      "openclaw_seconds": 2.360662939005124,
      "openclaw_ttft_seconds": 1.556977263004228,
      "latency_total_seconds": 32.30965048400685,
      "tool_call_detected": true,
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null,
      "provider": "Unsloth",
      "benchmark_track": "local"
    },
    {
      "model": "google/gemma-4-26b-a4b-it",
      "display_name": "Google: Gemma 4 26B A4B ",
      "run_status": "partial",
      "assistant_score": 65.229,
      "outcome": 54.123,
      "tool_use": 84.223,
      "grounding": 38.587,
      "state": 63.768,
      "english": 73.361,
      "safety": 94.901,
      "efficiency": 73.81,
      "tasks_passed": 11,
      "tasks_total": 21,
      "task_pass_rate": 52.38095238095238,
      "tool_call_success_rate": 97.6470588235294,
      "median_task_seconds": 9.68607979999797,
      "total_task_seconds": 202.65152759995544,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "deepseek/deepseek-v4-pro",
      "display_name": "DeepSeek: DeepSeek V4 Pro",
      "run_status": "partial",
      "assistant_score": 64.484,
      "outcome": 45.032,
      "tool_use": 88.592,
      "grounding": 57.609,
      "state": 54.106,
      "english": 73.361,
      "safety": 89.235,
      "efficiency": 59.524,
      "tasks_passed": 10,
      "tasks_total": 21,
      "task_pass_rate": 47.61904761904762,
      "tool_call_success_rate": 91.71974522292993,
      "median_task_seconds": 34.85148109997681,
      "total_task_seconds": 848.9442619999754,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-1m-mtp-q4-k-m",
      "display_name": "Qwythos-9B-Claude-Mythos-5-1M-MTP-Q4_K_M (Local)",
      "run_status": "partial",
      "assistant_score": 62.846,
      "outcome": 37.844,
      "tool_use": 85.437,
      "grounding": 49.457,
      "state": 58.937,
      "english": 83.607,
      "safety": 90.652,
      "efficiency": 78.571,
      "tasks_passed": 11,
      "tasks_total": 21,
      "task_pass_rate": 52.38095238095238,
      "tool_call_success_rate": 97.5609756097561,
      "median_task_seconds": 107.61787835699943,
      "total_task_seconds": 2291.1680853429934,
      "cold_start_seconds": 97.61175223300233,
      "cold_ttft_seconds": 0.0014789390043006279,
      "openclaw_seconds": 155.2012724169981,
      "openclaw_ttft_seconds": 0.03919271699851379,
      "latency_total_seconds": 252.81302465000044,
      "tool_call_detected": true,
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null,
      "provider": "Community local",
      "benchmark_track": "local"
    },
    {
      "model": "qwen/qwen3.7-plus",
      "display_name": "Qwen: Qwen3.7 Plus",
      "run_status": "partial",
      "assistant_score": 62.449,
      "outcome": 32.981,
      "tool_use": 85.437,
      "grounding": 63.043,
      "state": 63.768,
      "english": 69.262,
      "safety": 94.901,
      "efficiency": 73.81,
      "tasks_passed": 11,
      "tasks_total": 21,
      "task_pass_rate": 52.38095238095238,
      "tool_call_success_rate": 93.63636363636364,
      "median_task_seconds": 18.625667599975714,
      "total_task_seconds": 386.2151920999022,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "x-ai/grok-4.3",
      "display_name": "SpaceXAI: Grok 4.3",
      "run_status": "partial",
      "assistant_score": 60.728,
      "outcome": 37.209,
      "tool_use": 76.214,
      "grounding": 38.587,
      "state": 63.768,
      "english": 91.803,
      "safety": 100.0,
      "efficiency": 83.333,
      "tasks_passed": 11,
      "tasks_total": 21,
      "task_pass_rate": 52.38095238095238,
      "tool_call_success_rate": 98.6842105263158,
      "median_task_seconds": 8.675497100004577,
      "total_task_seconds": 182.00998060007987,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "qwen3.6-35b-hauhaucs-aggressive",
      "display_name": "Qwen3.6 35B A3B Uncensored HauhauCS Aggressive Q4_K_M (Local)",
      "run_status": "partial",
      "assistant_score": 60.682,
      "outcome": 43.975,
      "tool_use": 85.437,
      "grounding": 38.587,
      "state": 49.275,
      "english": 79.508,
      "safety": 84.986,
      "efficiency": 64.286,
      "tasks_passed": 9,
      "tasks_total": 21,
      "task_pass_rate": 42.857142857142854,
      "tool_call_success_rate": 96.46017699115045,
      "median_task_seconds": 9.014303698000731,
      "total_task_seconds": 216.09442760101228,
      "cold_start_seconds": 29.923409277005703,
      "cold_ttft_seconds": 0.003349618003994692,
      "openclaw_seconds": 2.1245510640001157,
      "openclaw_ttft_seconds": 1.4488617530005286,
      "latency_total_seconds": 32.04796034100582,
      "tool_call_detected": true,
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null,
      "provider": "Qwen fine-tunes",
      "benchmark_track": "local"
    },
    {
      "model": "stepfun/step-3.7-flash",
      "display_name": "StepFun: Step 3.7 Flash",
      "run_status": "partial",
      "assistant_score": 59.739,
      "outcome": 37.632,
      "tool_use": 84.466,
      "grounding": 46.739,
      "state": 39.614,
      "english": 73.361,
      "safety": 97.167,
      "efficiency": 83.333,
      "tasks_passed": 9,
      "tasks_total": 21,
      "task_pass_rate": 42.857142857142854,
      "tool_call_success_rate": 98.79518072289157,
      "median_task_seconds": 16.67475870002818,
      "total_task_seconds": 367.73295640001015,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "gemma-4-e4b-it",
      "display_name": "Gemma 4 E4B IT UD-Q4_K_XL (Local)",
      "run_status": "partial",
      "assistant_score": 58.785,
      "outcome": 34.038,
      "tool_use": 64.563,
      "grounding": 46.739,
      "state": 68.599,
      "english": 91.803,
      "safety": 97.167,
      "efficiency": 90.476,
      "tasks_passed": 9,
      "tasks_total": 21,
      "task_pass_rate": 42.857142857142854,
      "tool_call_success_rate": 94.73684210526316,
      "median_task_seconds": 4.518554032001703,
      "total_task_seconds": 110.72283329803031,
      "cold_start_seconds": 9.733592305005004,
      "cold_ttft_seconds": 0.005922943004406989,
      "openclaw_seconds": 1.552729579998413,
      "openclaw_ttft_seconds": 0.7556925419994514,
      "latency_total_seconds": 11.286321885003417,
      "tool_call_detected": true,
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null,
      "provider": "Google local",
      "benchmark_track": "local"
    },
    {
      "model": "nvidia/nemotron-3-ultra-550b-a55b",
      "display_name": "NVIDIA: Nemotron 3 Ultra",
      "run_status": "partial",
      "assistant_score": 58.648,
      "outcome": 44.186,
      "tool_use": 84.951,
      "grounding": 27.717,
      "state": 51.691,
      "english": 65.164,
      "safety": 97.167,
      "efficiency": 69.048,
      "tasks_passed": 9,
      "tasks_total": 21,
      "task_pass_rate": 42.857142857142854,
      "tool_call_success_rate": 99.02912621359224,
      "median_task_seconds": 10.869041700017988,
      "total_task_seconds": 239.14691919994948,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "google/gemini-3.5-flash-lite",
      "display_name": "Google: Gemini 3.5 Flash Lite",
      "run_status": "ok",
      "assistant_score": 57.38,
      "outcome": 47.78,
      "tool_use": 69.175,
      "grounding": 40.761,
      "state": 39.614,
      "english": 63.115,
      "safety": 92.068,
      "efficiency": 95.238,
      "tasks_passed": 10,
      "tasks_total": 21,
      "task_pass_rate": 47.61904761904762,
      "tool_call_success_rate": 100.0,
      "median_task_seconds": 2.658011000006809,
      "total_task_seconds": 68.38918980000017,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "google/gemini-3.6-flash",
      "display_name": "Google: Gemini 3.6 Flash",
      "run_status": "partial",
      "assistant_score": 56.269,
      "outcome": 29.81,
      "tool_use": 86.165,
      "grounding": 44.022,
      "state": 39.614,
      "english": 69.262,
      "safety": 92.068,
      "efficiency": 73.81,
      "tasks_passed": 7,
      "tasks_total": 21,
      "task_pass_rate": 33.333333333333336,
      "tool_call_success_rate": 98.0,
      "median_task_seconds": 9.219625000012456,
      "total_task_seconds": 205.46809570003825,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "openai/gpt-4.1",
      "display_name": "OpenAI: GPT-4.1",
      "run_status": "partial",
      "assistant_score": 53.072,
      "outcome": 28.33,
      "tool_use": 67.718,
      "grounding": 27.717,
      "state": 51.691,
      "english": 93.852,
      "safety": 92.918,
      "efficiency": 85.714,
      "tasks_passed": 6,
      "tasks_total": 21,
      "task_pass_rate": 28.571428571428573,
      "tool_call_success_rate": 94.44444444444444,
      "median_task_seconds": 5.031646200004616,
      "total_task_seconds": 107.87679090007441,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "deepseek/deepseek-v3.2",
      "display_name": "DeepSeek: DeepSeek V3.2",
      "run_status": "partial",
      "assistant_score": 53.054,
      "outcome": 36.152,
      "tool_use": 81.796,
      "grounding": 19.022,
      "state": 63.768,
      "english": 46.721,
      "safety": 100.0,
      "efficiency": 57.143,
      "tasks_passed": 6,
      "tasks_total": 21,
      "task_pass_rate": 28.571428571428573,
      "tool_call_success_rate": 97.02970297029702,
      "median_task_seconds": 19.721056299982592,
      "total_task_seconds": 420.250020600055,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "gemma-4-12b-obliterated",
      "display_name": "Gemma 4 12B Obliterated Q4_K_M (Local)",
      "run_status": "ok",
      "assistant_score": 47.075,
      "outcome": 18.393,
      "tool_use": 68.204,
      "grounding": 35.87,
      "state": 39.614,
      "english": 59.016,
      "safety": 97.167,
      "efficiency": 88.095,
      "tasks_passed": 6,
      "tasks_total": 21,
      "task_pass_rate": 28.571428571428573,
      "tool_call_success_rate": 91.17647058823529,
      "median_task_seconds": 11.005770889001724,
      "total_task_seconds": 222.64030131098843,
      "cold_start_seconds": 11.65942994600482,
      "cold_ttft_seconds": 0.004682411999965552,
      "openclaw_seconds": 2.4915786320052575,
      "openclaw_ttft_seconds": 0.9312468170028296,
      "latency_total_seconds": 14.151008578010078,
      "tool_call_detected": true,
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null,
      "provider": "Google local",
      "benchmark_track": "local"
    },
    {
      "model": "cydonia-24b-v4.3",
      "display_name": "Cydonia 24B v4.3 Q4_K_M (Local)",
      "run_status": "ok",
      "assistant_score": 33.93,
      "outcome": 16.913,
      "tool_use": 10.922,
      "grounding": 27.717,
      "state": 39.614,
      "english": 83.607,
      "safety": 92.918,
      "efficiency": 100.0,
      "tasks_passed": 3,
      "tasks_total": 21,
      "task_pass_rate": 14.285714285714286,
      "tool_call_success_rate": 100.0,
      "median_task_seconds": 24.692161588995077,
      "total_task_seconds": 858.8453925090071,
      "cold_start_seconds": 27.967421191999165,
      "cold_ttft_seconds": 0.005088760997750796,
      "openclaw_seconds": 9.215588352999475,
      "openclaw_ttft_seconds": 2.350872294999135,
      "latency_total_seconds": 37.18300954499864,
      "tool_call_detected": false,
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null,
      "provider": "Community local",
      "benchmark_track": "local"
    },
    {
      "model": "dolphin-mistral-24b-venice",
      "display_name": "Dolphin Mistral 24B Venice Edition Q4_K_M (Local)",
      "run_status": "ok",
      "assistant_score": 32.166,
      "outcome": 13.742,
      "tool_use": 10.922,
      "grounding": 19.565,
      "state": 39.614,
      "english": 87.705,
      "safety": 92.918,
      "efficiency": 100.0,
      "tasks_passed": 2,
      "tasks_total": 21,
      "task_pass_rate": 9.523809523809524,
      "tool_call_success_rate": 100.0,
      "median_task_seconds": 1.914171438998892,
      "total_task_seconds": 65.07358583799942,
      "cold_start_seconds": 17.65202158899774,
      "cold_ttft_seconds": 0.0032104339989018627,
      "openclaw_seconds": 1.020600588999514,
      "openclaw_ttft_seconds": 0.6275144590035779,
      "latency_total_seconds": 18.672622177997255,
      "tool_call_detected": false,
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null,
      "provider": "Community local",
      "benchmark_track": "local"
    },
    {
      "model": "google/gemma-3-12b-it",
      "display_name": "Google: Gemma 3 12B",
      "run_status": "ok",
      "assistant_score": 31.872,
      "outcome": 9.514,
      "tool_use": 10.922,
      "grounding": 19.565,
      "state": 39.614,
      "english": 100.0,
      "safety": 87.819,
      "efficiency": 100.0,
      "tasks_passed": 2,
      "tasks_total": 21,
      "task_pass_rate": 9.523809523809524,
      "tool_call_success_rate": 100.0,
      "median_task_seconds": 3.862632999997004,
      "total_task_seconds": 68.39876309997635,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    },
    {
      "model": "openai/gpt-5.6-sol",
      "display_name": "OpenAI: GPT-5.6 Sol",
      "run_status": "partial",
      "assistant_score": 15.775,
      "outcome": 15.222,
      "tool_use": 14.32,
      "grounding": 35.326,
      "state": 0.0,
      "english": 12.295,
      "safety": 12.465,
      "efficiency": 9.524,
      "tasks_passed": 3,
      "tasks_total": 21,
      "task_pass_rate": 14.285714285714286,
      "tool_call_success_rate": 100.0,
      "median_task_seconds": 0.0,
      "total_task_seconds": 40.589053800009424,
      "cold_start_seconds": 0.0,
      "cold_ttft_seconds": 0.0,
      "openclaw_seconds": 0.0,
      "openclaw_ttft_seconds": 0.0,
      "latency_total_seconds": 0.0,
      "tool_call_detected": false,
      "provider": "OpenRouter",
      "benchmark_track": "openrouter",
      "error": "",
      "provider_reported_cost_usd": null,
      "provider_cost_limit_usd": null
    }
  ]
}
