Agentic AI Atlas

II.

EvalResult JSON

eval-result:gpqa-diamond.gpt-5-4-mini.2026-03-17.accuracy

Structured · live

eval-result:gpqa-diamond.gpt-5-4-mini.2026-03-17.accuracy json

Inspect the normalized record payload exactly as the atlas UI reads it.

File · benchmarks/eval-results/eval-results-openai.yamlCluster · benchmarks

Record JSON

{
  "id": "eval-result:gpqa-diamond.gpt-5-4-mini.2026-03-17.accuracy",
  "_kind": "EvalResult",
  "_file": "benchmarks/eval-results/eval-results-openai.yaml",
  "_cluster": "benchmarks",
  "attributes": {
    "evalRunId": "eval-run:gpqa-diamond.gpt-5-4-mini.2026-03-17",
    "metricName": "accuracy",
    "score": 0.88,
    "unit": "fraction",
    "passFail": "pass",
    "reportedAt": "2026-03-17T00:00:00Z"
  },
  "outgoingEdges": [
    {
      "from": "eval-result:gpqa-diamond.gpt-5-4-mini.2026-03-17.accuracy",
      "to": "eval-run:gpqa-diamond.gpt-5-4-mini.2026-03-17",
      "kind": "belongs_to_eval_run"
    },
    {
      "from": "eval-result:gpqa-diamond.gpt-5-4-mini.2026-03-17.accuracy",
      "to": "benchmark:gpqa",
      "kind": "scored_against",
      "attributes": {}
    }
  ],
  "incomingEdges": []
}