judge:exact-match

{
  "id": "judge:exact-match",
  "_kind": "Judge",
  "_file": "benchmarks/eval-harnesses/judges.yaml",
  "_cluster": "benchmarks",
  "attributes": {
    "displayName": "Exact-match programmatic judge",
    "judgeKind": "programmatic",
    "notes": "Deterministic exact-match judge for closed-form benchmarks (e.g. MMLU).\n"
  },
  "outgoingEdges": [],
  "incomingEdges": [
    {
      "from": "eval-run:gaia.claude-code.2025",
      "to": "judge:exact-match",
      "kind": "judged_by",
      "attributes": {}
    }
  ]
}