[
  {
    "benchmark": "swe-bench-multimodal",
    "score": 35.9,
    "metric": "solveable_accuracy",
    "cost_per_instance": 2.97,
    "average_runtime": 1434.0,
    "full_archive": "https://results.eval.all-hands.dev/eval-21383028167-gpt-5-2-co_litellm_proxy-gpt-5-2-codex_26-01-27-09-13.tar.gz",
    "tags": [
      "swe-bench-multimodal"
    ],
    "agent_version": "v1.8.3",
    "submission_time": "2026-01-26T23:24:40.642068+00:00",
    "component_scores": {
      "solveable_accuracy": 35.9,
      "unsolveable_accuracy": 0.0,
      "combined_accuracy": 24.0,
      "solveable_resolved": 23,
      "solveable_total": 64,
      "unsolveable_resolved": 0,
      "unsolveable_total": 32
    }
  },
  {
    "benchmark": "gaia",
    "score": 70.9,
    "metric": "accuracy",
    "cost_per_instance": 0.55,
    "average_runtime": 799.0,
    "full_archive": "https://results.eval.all-hands.dev/eval-21377864374-gpt-5-2-co_litellm_proxy-gpt-5-2-codex_26-01-27-04-20.tar.gz",
    "tags": [
      "gaia"
    ],
    "agent_version": "v1.8.3",
    "submission_time": "2026-01-26T23:24:40.642068+00:00"
  },
  {
    "benchmark": "swe-bench",
    "score": 73.8,
    "metric": "accuracy",
    "cost_per_instance": 0.94,
    "average_runtime": 438.0,
    "full_archive": "https://results.eval.all-hands.dev/eval-21386738547-gpt-5-2-co_litellm_proxy-gpt-5-2-codex_26-01-27-12-57.tar.gz",
    "tags": [
      "swe-bench"
    ],
    "agent_version": "v1.8.3",
    "submission_time": "2026-01-26T23:24:40.642068+00:00"
  },
  {
    "benchmark": "commit0",
    "score": 43.8,
    "metric": "accuracy",
    "cost_per_instance": 5.5,
    "average_runtime": 1559.0,
    "full_archive": "https://results.eval.all-hands.dev/commit0/litellm_proxy-gpt-5-2-codex/24787401776/results.tar.gz",
    "tags": [
      "commit0"
    ],
    "agent_version": "v1.17.0",
    "submission_time": "2026-04-22T20:22:24+00:00",
    "eval_visualization_page": "https://laminar.sh/shared/evals/57c6395a-4cd4-4346-a0c0-4b1209275523"
  },
  {
    "benchmark": "swt-bench",
    "score": 67.0,
    "metric": "accuracy",
    "cost_per_instance": 0.66,
    "average_runtime": 344.0,
    "full_archive": "https://results.eval.all-hands.dev/swtbench/litellm_proxy-gpt-5-2-codex/23816423289/results.tar.gz",
    "tags": [
      "swt-bench"
    ],
    "agent_version": "v1.8.3",
    "submission_time": "2026-03-31T22:17:38+00:00"
  }
]