{
  "aliases": [],
  "authors": [],
  "details": {
    "benchmark_protocol": {
      "actions": [
        "Android operations through the agent framework"
      ],
      "counts": [
        {
          "basis": "reported",
          "evidence_url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md",
          "population": "benchmark task collection",
          "qualifier": "exact",
          "unit": "tasks",
          "value": 138
        },
        {
          "basis": "reported",
          "evidence_url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md",
          "population": "benchmark application set",
          "qualifier": "exact",
          "unit": "applications",
          "value": 9
        }
      ],
      "evaluation": "The evaluation pipeline produces success, subgoal success, reversed redundancy, and reasonable-operation metrics; result generation requires a configured judge model.",
      "evidence_urls": [
        "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md"
      ],
      "focus": [
        "Android task completion",
        "subgoal completion",
        "operation efficiency"
      ],
      "notes": "The benchmark uses predefined virtual-device images and apps chosen to work without network/login requirements. XML and SoM modes should be reported separately. These counts concern evaluation tasks, not the separately released Android Instruct training set.",
      "observations": [
        "UI XML",
        "set-of-marks views"
      ],
      "setting": "interactive"
    },
    "evaluation_type": "benchmark task evaluation",
    "leaderboard_url": "https://docs.google.com/spreadsheets/d/1Zv6mBfd4Ibt8mke24K6zAFrXe4AvEFaBLW3hpJqSgjw/edit?gid=0#gid=0",
    "metrics": [
      "task success"
    ],
    "task_count": 138
  },
  "environments": [
    "android"
  ],
  "evidence": [
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-16T06:35:58Z",
      "role": "primary",
      "sha256": "340680a9a7833d103234b8367195deebd29bfec7e25ebc50c2dc04fc3627e55b",
      "supports": [
        "name",
        "kind",
        "summary",
        "environments",
        "topics",
        "scope",
        "relations",
        "details"
      ],
      "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md"
    },
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-16T06:35:58Z",
      "role": "primary",
      "sha256": "340680a9a7833d103234b8367195deebd29bfec7e25ebc50c2dc04fc3627e55b",
      "supports": [
        "details.benchmark_protocol"
      ],
      "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md"
    },
    {
      "method": "license_header",
      "retrieved_at": "2026-10-16T06:41:58Z",
      "role": "primary",
      "sha256": "1bbbee919161337e2456c587c6be809214406482cd9f1292a9eb62428f6961f3",
      "supports": [
        "licenses.code"
      ],
      "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/LICENSE"
    }
  ],
  "id": "github:thudm/android-lab",
  "identifiers": {},
  "kind": "benchmark",
  "last_verified_at": "2026-10-16T06:35:58Z",
  "licenses": {
    "code": {
      "basis": "repository LICENSE header; not a weights/data license",
      "declared": "MIT",
      "retrieved_at": "2026-10-16T06:41:58Z",
      "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/LICENSE"
    },
    "data": null,
    "document": null,
    "weights": null
  },
  "links": [
    {
      "rel": "primary",
      "url": "https://github.com/THUDM/Android-Lab"
    }
  ],
  "name": "AndroidLab",
  "notes": "",
  "published_at": null,
  "relations": [
    {
      "evidence_url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md",
      "target_id": "arxiv:2410.24024",
      "type": "documented_by"
    }
  ],
  "review_status": "curated",
  "scope": "core",
  "summary": "Reproducible Android benchmark of 138 tasks across nine apps on predefined virtual devices, run via AVD on Mac or Docker on Linux, released with the Android Instruct training set.",
  "topics": [
    "evaluation",
    "mobile-navigation"
  ],
  "updated_at": null,
  "url": "https://github.com/THUDM/Android-Lab"
}