{
  "id": "benchmark:terminal-bench",
  "_kind": "Benchmark",
  "_file": "benchmarks/benchmarks/terminal-bench.yaml",
  "_cluster": "benchmarks",
  "attributes": {
    "displayName": "Terminal-Bench",
    "homepageUrl": "https://www.tbench.ai/",
    "kind": "terminal-agent",
    "targetsKind": "AgentVersion",
    "description": "Benchmark for AI agents performing real software-engineering and devops tasks in a terminal environment.\n"
  },
  "outgoingEdges": [
    {
      "from": "benchmark:terminal-bench",
      "to": "skill-area:cli-design",
      "kind": "covers",
      "attributes": {
        "attributes": {
          "coverage": "partial",
          "weight": 0.4
        }
      }
    }
  ],
  "incomingEdges": [
    {
      "from": "eval-result:terminal-bench.claude-sonnet-4-5.001",
      "to": "benchmark:terminal-bench",
      "kind": "scored_against",
      "attributes": {}
    },
    {
      "from": "eval-run:terminal-bench.claude-sonnet-4-5.2025-09",
      "to": "benchmark:terminal-bench",
      "kind": "for_benchmark"
    },
    {
      "from": "test-set:terminal-bench-v1",
      "to": "benchmark:terminal-bench",
      "kind": "belongs_to_benchmark"
    },
    {
      "from": "scope-boundary:terminal-bench.scope",
      "to": "benchmark:terminal-bench",
      "kind": "bounds_subject"
    }
  ]
}