{
  "scope": "Inference support only; model weights unchanged",
  "unit_tests": "PASS",
  "selected_cases": 9,
  "tests": [
    {
      "id": "MMLU-Pro:001",
      "thinking": false,
      "completed": true,
      "finish_reason": "eos",
      "output_tokens": 50,
      "seconds": 2.9331285804510117
    },
    {
      "id": "HumanEval-Plus:001",
      "thinking": false,
      "completed": true,
      "finish_reason": "eos",
      "output_tokens": 233,
      "seconds": 10.614582307636738
    },
    {
      "id": "MATH-Level-5:002",
      "thinking": false,
      "completed": true,
      "finish_reason": "eos",
      "output_tokens": 116,
      "seconds": 5.350349590182304
    },
    {
      "id": "GSM8K:001",
      "thinking": false,
      "completed": true,
      "finish_reason": "eos",
      "output_tokens": 111,
      "seconds": 5.108682680875063
    },
    {
      "id": "Q36-Hermes-Tool-Format:001",
      "thinking": false,
      "completed": true,
      "finish_reason": "eos",
      "output_tokens": 41,
      "seconds": 1.973704420030117
    },
    {
      "id": "Q36-Output-Integrity:001",
      "thinking": false,
      "completed": true,
      "finish_reason": "eos",
      "output_tokens": 2,
      "seconds": 0.2203267514705658
    },
    {
      "id": "MMLU-Pro:001",
      "thinking": true,
      "completed": false,
      "finish_reason": "repeated_block",
      "output_tokens": 432,
      "seconds": 19.451829474419355
    },
    {
      "id": "HumanEval-Plus:001",
      "thinking": true,
      "completed": false,
      "finish_reason": "time_limit",
      "output_tokens": 1005,
      "seconds": 45.0059671998024
    },
    {
      "id": "MATH-Level-5:002",
      "thinking": true,
      "completed": false,
      "finish_reason": "repeated_block",
      "output_tokens": 352,
      "seconds": 15.891684971749783
    }
  ],
  "limitations": "Selected known failures plus regressions; not independent benchmark accuracy; raw diagnostic outputs are kept privately."
}