[
  {
    "test": "ifeval",
    "thinking": "on",
    "status": "success",
    "score": 0.9186,
    "main_metric": "final_acc",
    "metrics": {
      "prompt_strict_acc": 0.9,
      "prompt_strict_stderr": 0.0302,
      "prompt_loose_acc": 0.91,
      "prompt_loose_stderr": 0.0288,
      "inst_strict_acc": 0.929,
      "inst_strict_stderr": 0.0223,
      "inst_loose_acc": 0.9355,
      "inst_loose_stderr": 0.0215,
      "final_acc": 0.9186,
      "final_stderr": 0.0317
    },
    "samples": 100,
    "epochs": 1,
    "input_tokens": 9857,
    "output_tokens": 147729,
    "reasoning_tokens": 0,
    "seconds": 961.4,
    "error": null,
    "log_file": "inspect/thinking_on/2026-10-19T00-53-29-00-00_ifeval_nBYADGwpFEoaTm46TbuGwZ.eval",
    "hidden_prompt": {
      "requests": 100,
      "flagged": 0,
      "tokens_per_char": 0.3431,
      "max_excess_tokens": -391,
      "examples": [],
      "error": null
    }
  },
  {
    "test": "mmlu_pro",
    "thinking": "on",
    "status": "success",
    "score": 0.9,
    "main_metric": "accuracy",
    "metrics": {
      "accuracy": 0.9,
      "stderr": 0.0429
    },
    "samples": 50,
    "epochs": 1,
    "input_tokens": 15784,
    "output_tokens": 192476,
    "reasoning_tokens": 0,
    "seconds": 1234.5,
    "error": null,
    "log_file": "inspect/thinking_on/2026-10-19T01-09-30-00-00_mmlu-pro_VtTgypNpFM4nSEE53J5wTc.eval",
    "hidden_prompt": {
      "requests": 50,
      "flagged": 0,
      "tokens_per_char": 0.2928,
      "max_excess_tokens": -561,
      "examples": [],
      "error": null
    }
  },
  {
    "test": "livecodebench_pro",
    "thinking": "on",
    "status": "success",
    "score": 0.8333,
    "main_metric": "accuracy",
    "metrics": {
      "accuracy": 0.8333,
      "stderr": 0.0692
    },
    "samples": 30,
    "epochs": 1,
    "input_tokens": 24747,
    "output_tokens": 263456,
    "reasoning_tokens": 0,
    "seconds": 1949.7,
    "error": null,
    "log_file": "inspect/thinking_on/2026-10-19T01-30-16-00-00_livecodebench-pro_Cy6qCHoUxLy5FNFMCwYZvE.eval",
    "hidden_prompt": {
      "requests": 30,
      "flagged": 0,
      "tokens_per_char": 0.3166,
      "max_excess_tokens": -1489,
      "examples": [],
      "error": null
    }
  },
  {
    "test": "bfcl",
    "thinking": "on",
    "status": "success",
    "score": 0.67,
    "main_metric": "accuracy",
    "metrics": {
      "accuracy": 0.67,
      "stderr": 0.0473,
      "multi_turn_long_context_acc": 0.6,
      "live_multiple_acc": 0.7619,
      "exec_parallel_multiple_acc": 1.0,
      "live_irrelevance_acc": 0.7647,
      "multi_turn_composite_acc": 0.0,
      "exec_parallel_acc": 0.75,
      "live_simple_acc": 0.8571,
      "simple_python_acc": 0.8182,
      "multiple_acc": 1.0,
      "parallel_multiple_acc": 1.0,
      "irrelevance_acc": 0.5,
      "sql_acc": 0.0,
      "multi_turn_miss_func_acc": 0.25,
      "multi_turn_base_acc": 1.0,
      "multi_turn_miss_param_acc": 0.0,
      "live_parallel_multiple_acc": 0.0,
      "simple_java_acc": 0.0,
      "all": 0.67,
      "multi_turn_long_context_sterr": 0.2449,
      "live_multiple_sterr": 0.0952,
      "exec_parallel_multiple_sterr": 0.0,
      "live_irrelevance_sterr": 0.106,
      "multi_turn_composite_sterr": 0.0,
      "exec_parallel_sterr": 0.25,
      "live_simple_sterr": 0.1429,
      "simple_python_sterr": 0.122,
      "multiple_sterr": 0.0,
      "parallel_multiple_sterr": 0.0,
      "irrelevance_sterr": 0.2887,
      "sql_sterr": 0.0,
      "multi_turn_miss_func_sterr": 0.1637,
      "multi_turn_base_sterr": 0.0,
      "multi_turn_miss_param_sterr": 0.0,
      "live_parallel_multiple_sterr": 0.0,
      "simple_java_sterr": 0.0,
      "all2": 0.0473
    },
    "samples": 100,
    "epochs": 1,
    "input_tokens": 313719,
    "output_tokens": 122010,
    "reasoning_tokens": 0,
    "seconds": 1225.2,
    "error": null,
    "log_file": "inspect/thinking_on/2026-10-19T02-02-35-00-00_bfcl_4Rc6mz8sM3udu8iym8tEJx.eval",
    "hidden_prompt": {
      "requests": 405,
      "flagged": 0,
      "tokens_per_char": 0.2393,
      "max_excess_tokens": -431,
      "examples": [],
      "error": null
    }
  },
  {
    "test": "ifeval",
    "thinking": "off",
    "status": "success",
    "score": 0.8898,
    "main_metric": "final_acc",
    "metrics": {
      "prompt_strict_acc": 0.85,
      "prompt_strict_stderr": 0.0359,
      "prompt_loose_acc": 0.89,
      "prompt_loose_stderr": 0.0314,
      "inst_strict_acc": 0.8968,
      "inst_strict_stderr": 0.0243,
      "inst_loose_acc": 0.9226,
      "inst_loose_stderr": 0.0222,
      "final_acc": 0.8898,
      "final_stderr": 0.0372
    },
    "samples": 100,
    "epochs": 1,
    "input_tokens": 5857,
    "output_tokens": 27369,
    "reasoning_tokens": 0,
    "seconds": 257.3,
    "error": null,
    "log_file": "inspect/thinking_off/2026-10-19T02-23-00-00-00_ifeval_YWCaRmizzgt6ye894pNUG2.eval",
    "hidden_prompt": {
      "requests": 100,
      "flagged": 0,
      "tokens_per_char": 0.1919,
      "max_excess_tokens": -431,
      "examples": [],
      "error": null
    }
  },
  {
    "test": "mmlu_pro",
    "thinking": "off",
    "status": "success",
    "score": 0.82,
    "main_metric": "accuracy",
    "metrics": {
      "accuracy": 0.82,
      "stderr": 0.0549
    },
    "samples": 50,
    "epochs": 1,
    "input_tokens": 13784,
    "output_tokens": 58589,
    "reasoning_tokens": 0,
    "seconds": 338.9,
    "error": null,
    "log_file": "inspect/thinking_off/2026-10-19T02-27-18-00-00_mmlu-pro_dXAPVDQsoopypwjxJY3REQ.eval",
    "hidden_prompt": {
      "requests": 50,
      "flagged": 0,
      "tokens_per_char": 0.2475,
      "max_excess_tokens": -601,
      "examples": [],
      "error": null
    }
  },
  {
    "test": "livecodebench_pro",
    "thinking": "off",
    "status": "success",
    "score": 0.5,
    "main_metric": "accuracy",
    "metrics": {
      "accuracy": 0.5,
      "stderr": 0.0928
    },
    "samples": 30,
    "epochs": 1,
    "input_tokens": 23667,
    "output_tokens": 217232,
    "reasoning_tokens": 0,
    "seconds": 1517.0,
    "error": null,
    "log_file": "inspect/thinking_off/2026-10-19T02-33-08-00-00_livecodebench-pro_gZZ774nokMiLjZ8UVzDE5F.eval",
    "hidden_prompt": {
      "requests": 30,
      "flagged": 0,
      "tokens_per_char": 0.3009,
      "max_excess_tokens": -1525,
      "examples": [],
      "error": null
    }
  },
  {
    "test": "bfcl",
    "thinking": "off",
    "status": "success",
    "score": 0.57,
    "main_metric": "accuracy",
    "metrics": {
      "accuracy": 0.57,
      "stderr": 0.0498,
      "multi_turn_long_context_acc": 0.8,
      "live_multiple_acc": 0.7143,
      "exec_parallel_multiple_acc": 0.5,
      "live_irrelevance_acc": 0.4706,
      "multi_turn_composite_acc": 0.0,
      "exec_parallel_acc": 0.25,
      "live_simple_acc": 0.8571,
      "simple_python_acc": 0.8182,
      "multiple_acc": 1.0,
      "parallel_multiple_acc": 0.6667,
      "irrelevance_acc": 0.25,
      "sql_acc": 0.0,
      "multi_turn_miss_func_acc": 0.375,
      "multi_turn_base_acc": 0.8,
      "multi_turn_miss_param_acc": 0.0,
      "live_parallel_multiple_acc": 0.0,
      "simple_java_acc": 0.0,
      "all": 0.57,
      "multi_turn_long_context_sterr": 0.2,
      "live_multiple_sterr": 0.101,
      "exec_parallel_multiple_sterr": 0.5,
      "live_irrelevance_sterr": 0.1248,
      "multi_turn_composite_sterr": 0.0,
      "exec_parallel_sterr": 0.25,
      "live_simple_sterr": 0.1429,
      "simple_python_sterr": 0.122,
      "multiple_sterr": 0.0,
      "parallel_multiple_sterr": 0.3333,
      "irrelevance_sterr": 0.25,
      "sql_sterr": 0.0,
      "multi_turn_miss_func_sterr": 0.183,
      "multi_turn_base_sterr": 0.2,
      "multi_turn_miss_param_sterr": 0.0,
      "live_parallel_multiple_sterr": 0.0,
      "simple_java_sterr": 0.0,
      "all2": 0.0498
    },
    "samples": 100,
    "epochs": 1,
    "input_tokens": 282762,
    "output_tokens": 32527,
    "reasoning_tokens": 0,
    "seconds": 720.3,
    "error": null,
    "log_file": "inspect/thinking_off/2026-10-19T02-58-15-00-00_bfcl_7jvSpcb9w3Cv2Q5urCkxcG.eval",
    "hidden_prompt": {
      "requests": 425,
      "flagged": 0,
      "tokens_per_char": 0.2473,
      "max_excess_tokens": -467,
      "examples": [],
      "error": null
    }
  }
]