iiRecord
Agentic AI Atlas · MedQA
benchmark:medqaa5c.ai
II.
Benchmark JSON

benchmark:medqa

Structured · live

MedQA json

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

File · benchmarks/benchmarks/benchmarks-reasoning.yamlCluster · benchmarks
Record JSON
{
  "id": "benchmark:medqa",
  "_kind": "Benchmark",
  "_file": "benchmarks/benchmarks/benchmarks-reasoning.yaml",
  "_cluster": "benchmarks",
  "attributes": {
    "displayName": "MedQA",
    "homepageUrl": "https://github.com/jind11/MedQA",
    "kind": "reasoning",
    "targetsKind": "ModelVersion",
    "description": "MedQA is a multiple-choice medical question-answering benchmark\ndrawn from US Medical Licensing Exam (USMLE) and Chinese / Taiwanese\nequivalents — the standard reference for medical-domain LLM eval.\n"
  },
  "outgoingEdges": [
    {
      "from": "benchmark:medqa",
      "to": "skill-area:medical-agent",
      "kind": "covers",
      "attributes": {}
    },
    {
      "from": "benchmark:medqa",
      "to": "domain:healthcare",
      "kind": "applies_to",
      "attributes": {}
    }
  ],
  "incomingEdges": []
}