{
  "aliases": [],
  "authors": [],
  "details": {
    "benchmark_protocol": {
      "actions": [
        "click coordinates"
      ],
      "counts": [],
      "evaluation": "The inspected Jedi evaluation adapter reads the v2 mobile/desktop/web annotation files, scores a point against the target box, and reports overall, platform, and text/icon accuracy. This is a named compatible evaluator implementation; the OS-Atlas dataset card itself supplies no evaluator specification.",
      "evidence_urls": [
        "https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2",
        "https://osatlas.github.io/",
        "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_screenspot_v2_jedi.py",
        "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"
      ],
      "focus": [
        "single-step GUI grounding on the ScreenSpot-v2 release"
      ],
      "notes": "The public Hub revision 5efbb1f1 contains three platform JSON annotation files and an image ZIP; its card only declares Apache-2.0. No sample total or equivalence to the original ScreenSpot test count is asserted. Keep v2 separate from v1 and retain evaluator/version and coordinate conversion when comparing results. Files were listed, not downloaded as data.",
      "observations": [
        "screenshot",
        "instruction"
      ],
      "setting": "offline"
    },
    "disabled": false,
    "download_url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2/tree/main",
    "gated": false,
    "hub_id": "OS-Copilot/ScreenSpot-v2",
    "languages": [],
    "modalities": [
      "image"
    ],
    "size_category_reported": null
  },
  "environments": [
    "cross-platform"
  ],
  "evidence": [
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-16T06:38:26Z",
      "role": "primary",
      "sha256": "7feddadc306e384f8df6a3d9c8afc0c776cb3734c4c7fe3eb842eab85537f408",
      "supports": [
        "name",
        "kind",
        "summary",
        "environments",
        "topics",
        "scope",
        "relations",
        "details.benchmark_protocol"
      ],
      "url": "https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2"
    },
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-17T14:22:05Z",
      "role": "primary",
      "sha256": "94547cb88756cf6a36bfb86d5fe9bdca1e40d7edb9e452c33cb462bbe760fed0",
      "supports": [
        "details.benchmark_protocol"
      ],
      "url": "https://osatlas.github.io/"
    },
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-17T14:21:09Z",
      "role": "primary",
      "sha256": "0e8772bc8fd5a25c50d3c2f6b0caee05ed2ab4659d46d99d60adf2b5ab29e997",
      "supports": [
        "details.benchmark_protocol"
      ],
      "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_screenspot_v2_jedi.py"
    },
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-17T14:21:09Z",
      "role": "primary",
      "sha256": "b85f5685ce6c200ea648471666177859693d7f40cf50c12d08765ae68580de4a",
      "supports": [
        "details.benchmark_protocol"
      ],
      "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"
    },
    {
      "method": "huggingface_api",
      "retrieved_at": "2026-10-16T06:38:26Z",
      "role": "primary",
      "sha256": "7feddadc306e384f8df6a3d9c8afc0c776cb3734c4c7fe3eb842eab85537f408",
      "supports": [
        "details",
        "licenses.data"
      ],
      "url": "https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2"
    }
  ],
  "id": "url:bd7214de6d98917c2350",
  "identifiers": {},
  "kind": "benchmark",
  "last_verified_at": "2026-10-16T06:38:26Z",
  "licenses": {
    "code": null,
    "data": {
      "basis": "publisher model/dataset card",
      "declared": "apache-2.0",
      "retrieved_at": "2026-10-16T06:38:26Z",
      "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"
    },
    "document": null,
    "weights": null
  },
  "links": [
    {
      "rel": "primary",
      "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"
    }
  ],
  "name": "ScreenSpot v2",
  "notes": "",
  "published_at": null,
  "relations": [],
  "review_status": "curated",
  "scope": "core",
  "summary": "Second version of the ScreenSpot grounding test set, published by OS-Copilot as separate desktop, mobile and web annotation files with a shared screenshot archive.",
  "topics": [
    "grounding",
    "evaluation"
  ],
  "updated_at": null,
  "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"
}