{
  "tag": "full",
  "n_tasks": 100,
  "k": 8,
  "budgets": [
    "no_think",
    "256",
    "512",
    "1024",
    "2048",
    "unbudgeted"
  ],
  "controls": false,
  "model": "Qwen/Qwen3.5-4B",
  "gen_secs": 31110.0,
  "conditions": {
    "no_think": {
      "n_tasks": 100,
      "greedy_pass@1": 0.76,
      "sampled_pass@1": 0.7762,
      "pass@8_oracle": 0.91,
      "visible_selector@1": 0.84,
      "oracle_minus_deployable": 0.07,
      "mean_think_tokens": 0,
      "mean_total_tokens": 121.3,
      "forced_close_frac": 0.0
    },
    "think_256": {
      "n_tasks": 100,
      "greedy_pass@1": 0.87,
      "sampled_pass@1": 0.8325,
      "pass@8_oracle": 0.93,
      "visible_selector@1": 0.88,
      "oracle_minus_deployable": 0.05,
      "mean_think_tokens": 246.4,
      "mean_total_tokens": 337.2,
      "forced_close_frac": 0.854
    },
    "think_512": {
      "n_tasks": 100,
      "greedy_pass@1": 0.87,
      "sampled_pass@1": 0.8575,
      "pass@8_oracle": 0.94,
      "visible_selector@1": 0.89,
      "oracle_minus_deployable": 0.05,
      "mean_think_tokens": 408.4,
      "mean_total_tokens": 494.2,
      "forced_close_frac": 0.422
    },
    "think_1024": {
      "n_tasks": 100,
      "greedy_pass@1": 0.91,
      "sampled_pass@1": 0.8788,
      "pass@8_oracle": 0.96,
      "visible_selector@1": 0.91,
      "oracle_minus_deployable": 0.05,
      "mean_think_tokens": 529.7,
      "mean_total_tokens": 604.7,
      "forced_close_frac": 0.131
    },
    "think_2048": {
      "n_tasks": 100,
      "greedy_pass@1": 0.86,
      "sampled_pass@1": 0.8638,
      "pass@8_oracle": 0.96,
      "visible_selector@1": 0.92,
      "oracle_minus_deployable": 0.04,
      "mean_think_tokens": 629.2,
      "mean_total_tokens": 706.2,
      "forced_close_frac": 0.074
    },
    "think_unbudgeted": {
      "n_tasks": 100,
      "greedy_pass@1": 0.84,
      "sampled_pass@1": 0.8638,
      "pass@8_oracle": 0.95,
      "visible_selector@1": 0.92,
      "oracle_minus_deployable": 0.03,
      "mean_think_tokens": 572.4,
      "mean_total_tokens": 810.9,
      "forced_close_frac": 0.0
    }
  }
}