{
  "id": "benchmark:assistant-bench",
  "_kind": "Benchmark",
  "_file": "benchmarks/benchmarks/benchmarks-agentic.yaml",
  "_cluster": "benchmarks",
  "attributes": {
    "displayName": "AssistantBench",
    "homepageUrl": "https://assistantbench.github.io/",
    "kind": "web-agent",
    "targetsKind": "AgentVersion",
    "description": "Real-world web-research tasks (price comparison, scheduling,\nfinding multi-attribute info) constructed to stress-test agentic\nassistants beyond simple QA.\n"
  },
  "outgoingEdges": [
    {
      "from": "benchmark:assistant-bench",
      "to": "skill-area:deep-web-research",
      "kind": "covers",
      "attributes": {
        "attributes": {
          "coverage": "partial",
          "weight": 0.4
        }
      }
    },
    {
      "from": "benchmark:assistant-bench",
      "to": "skill-area:agentic-loops",
      "kind": "covers",
      "attributes": {
        "attributes": {
          "coverage": "partial",
          "weight": 0.2
        }
      }
    },
    {
      "from": "benchmark:assistant-bench",
      "to": "domain:web-development",
      "kind": "applies_to",
      "attributes": {}
    }
  ],
  "incomingEdges": []
}