eval_stats.json

{
  "run_id": "20260519-162229",
  "model": "qwen/qwen3.6-flash",
  "truncated": false,
  "resumed": 0,
  "cost": {
    "calls": 41,
    "prompt_tokens": 332060,
    "completion_tokens": 43651,
    "usd": 0.0,
    "max_usd": 0.0
  },
  "summary": {
    "action-sequenced-execution:hard": {
      "n": 1,
      "win_rate": 0.0,
      "composite_mean": 0.1773,
      "composite_std": 0.0,
      "perception_mean": 0.6844,
      "reasoning_mean": 0.6737,
      "action_mean": 1.0,
      "objective_mean": 0.375,
      "weakest_link_hist": {
        "reasoning": 1
      }
    }
  },
  "overall": {
    "n": 1,
    "win_rate": 0.0,
    "composite_mean": 0.1773,
    "composite_std": 0.0,
    "perception_mean": 0.6844,
    "reasoning_mean": 0.6737,
    "action_mean": 1.0,
    "objective_mean": 0.375,
    "weakest_link_hist": {
      "reasoning": 1
    }
  },
  "reward_vector_mean": {
    "economy": 0.5,
    "military": 0.0,
    "territory": 0.5491,
    "scouting": 0.6,
    "objective": 0.375
  },
  "episodes": [
    {
      "cell": "action-sequenced-execution:hard",
      "capability": "action",
      "split": "public",
      "seed": 1,
      "outcome": "loss",
      "composite": 0.1773,
      "perception": 0.6844,
      "reasoning": 0.6737,
      "action": 1.0,
      "weakest_link": "reasoning",
      "objective_progress": 0.375,
      "reward_vector": {
        "economy": 0.5,
        "military": 0.0,
        "territory": 0.5491,
        "scouting": 0.6,
        "objective": 0.375
      },
      "turns": 41,
      "notes": [
        "objective not met (loss); weakest link: reasoning"
      ]
    }
  ],
  "skipped": []
}