{
 "schema_version": 1,
 "source": {
  "benchmark_scores_sha256": "baa9a1af5f18484be42a920194a1c6e940f2f91bbf86fb8c2721588886758aaa",
  "overall_resource_frontiers_sha256": "fed04f68399b48f9a0859ea79c9851d4003dba9b377ccfdd3df22b77dce85b0f",
  "full_results_sha256": "0689c57b37b5ddeb4ab91606a080ed320f57d294e87c85ceee84c95a04d5dfdd",
  "score_vs_tokens_sha256": "894a423bdf2a22b62ac58b8a710ac981315b3f179d80507ee913dba01574b896"
 },
 "policy": {
  "time": "Mean native agent_execution.finished_at minus started_at, in minutes; excludes environment/agent setup and independent final grading. Includes model/tool waiting; not active compute time.",
  "probes": "Mean published consumed evaluator probes (run.probe_count), including failed probes; independent final grading does not consume a probe. Not native tool-call counts.",
  "tokens": "Recorded harness input + output tokens over the complete selected run; input already includes cached input, which must not be added again"
 },
 "configs": {
  "codex:openrouter/openai/gpt-6-astra": {
   "design": {
    "tasks": 16,
    "minutes": 119.66086844791667,
    "probes": 23.25,
    "tokens_m": 24.1798268125,
    "token_runs": 16,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 119.65926566333333,
    "probes": 28.7,
    "tokens_m": 22.2124819,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 119.66025199230766,
    "probes": 25.346153846153847,
    "tokens_m": 23.423155692307695,
    "tokens_lower_bound": false
   }
  },
  "claude-code:openrouter/anthropic/claude-opus-5.5": {
   "design": {
    "tasks": 16,
    "minutes": 96.17268171979168,
    "probes": 21.0625,
    "tokens_m": 13.543090875,
    "token_runs": 16,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 85.090858745,
    "probes": 27.2,
    "tokens_m": 5.9054181,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 91.9104421141026,
    "probes": 23.423076923076923,
    "tokens_m": 10.605524423076925,
    "tokens_lower_bound": false
   }
  },
  "codex:openrouter/openai/gpt-6.1-sol": {
   "design": {
    "tasks": 16,
    "minutes": 119.6581875572917,
    "probes": 22.0625,
    "tokens_m": 21.215253,
    "token_runs": 16,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 119.65902141166666,
    "probes": 29.6,
    "tokens_m": 18.716333200000005,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 119.65850827051284,
    "probes": 24.96153846153846,
    "tokens_m": 20.25413,
    "tokens_lower_bound": false
   }
  },
  "claude-code:openrouter/anthropic/claude-fable-5.1": {
   "design": {
    "tasks": 16,
    "minutes": 95.84206845729167,
    "probes": 16.5625,
    "tokens_m": 15.37799775,
    "token_runs": 16,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 80.82884697333334,
    "probes": 27.1,
    "tokens_m": 5.469468200000001,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 90.06775250192308,
    "probes": 20.615384615384617,
    "tokens_m": 11.567024846153846,
    "tokens_lower_bound": false
   }
  },
  "claude-code:openrouter/anthropic/claude-sonnet-5.5": {
   "design": {
    "tasks": 16,
    "minutes": 95.41179710208334,
    "probes": 23.0,
    "tokens_m": 21.8664311875,
    "token_runs": 16,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 56.94191501333333,
    "probes": 27.1,
    "tokens_m": 3.6625455,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 80.61568860641026,
    "probes": 24.576923076923077,
    "tokens_m": 14.864936692307692,
    "tokens_lower_bound": false
   }
  },
  "mini-swe-agent:openrouter/google/gemini-4-argon": {
   "design": {
    "tasks": 16,
    "minutes": 120.002104590625,
    "probes": 8.3125,
    "tokens_m": 19.6221491875,
    "token_runs": 16,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 120.00473618166667,
    "probes": 7.3,
    "tokens_m": 7.0172943,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 120.00311674102562,
    "probes": 7.923076923076923,
    "tokens_m": 14.774128076923079,
    "tokens_lower_bound": false
   }
  },
  "muse-code:openrouter/meta/muse-wonder": {
   "design": {
    "tasks": 16,
    "minutes": 99.79803287916668,
    "probes": 24.0,
    "tokens_m": 42.6422553125,
    "token_runs": 16,
    "tokens_lower_bound": true
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 75.95807851166667,
    "probes": 29.1,
    "tokens_m": 8.1463265,
    "token_runs": 10,
    "tokens_lower_bound": true
   },
   "overall": {
    "tasks": 26,
    "minutes": 90.62881966089745,
    "probes": 25.96153846153846,
    "tokens_m": 29.374590384615384,
    "tokens_lower_bound": true
   }
  },
  "codex:openrouter/openai/gpt-5.6-sol": {
   "design": {
    "tasks": 16,
    "minutes": 119.65947537708334,
    "probes": 31.125,
    "tokens_m": 34.99838486666667,
    "token_runs": 15,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 119.6615317666667,
    "probes": 29.5,
    "tokens_m": 7.941782,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 119.66026629615385,
    "probes": 30.5,
    "tokens_m": 24.17574372,
    "tokens_lower_bound": false
   }
  },
  "claude-code:openrouter/moonshotai/kimi-k3.1-pro": {
   "design": {
    "tasks": 16,
    "minutes": 114.90352407708335,
    "probes": 20.125,
    "tokens_m": 28.8693303125,
    "token_runs": 16,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 120.07347727833334,
    "probes": 29.5,
    "tokens_m": 17.7410553,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 116.89196761602564,
    "probes": 23.73076923076923,
    "tokens_m": 24.58922453846154,
    "tokens_lower_bound": false
   }
  },
  "mini-swe-agent:openrouter/z-ai/glm-5.3-flash": {
   "design": {
    "tasks": 16,
    "minutes": 88.60528512916666,
    "probes": 13.25,
    "tokens_m": 48.2860541875,
    "token_runs": 16,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 92.56925549833335,
    "probes": 14.4,
    "tokens_m": 6.7168967,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 90.12988911730767,
    "probes": 13.692307692307692,
    "tokens_m": 32.297916692307695,
    "tokens_lower_bound": false
   }
  },
  "claude-code:openrouter/z-ai/glm-5.3-flash": {
   "design": {
    "tasks": 16,
    "minutes": 114.09203072500001,
    "probes": 4.625,
    "tokens_m": 8.0541615625,
    "token_runs": 16,
    "tokens_lower_bound": false
   },
   "reimplementation": {
    "tasks": 10,
    "minutes": 85.967508455,
    "probes": 18.6,
    "tokens_m": 14.8644438,
    "token_runs": 10,
    "tokens_lower_bound": false
   },
   "overall": {
    "tasks": 26,
    "minutes": 103.27490677500002,
    "probes": 10.0,
    "tokens_m": 10.673500884615384,
    "tokens_lower_bound": false
   }
  }
 }
}