{
  "aliases": [],
  "authors": [],
  "details": {
    "benchmark_protocol": {
      "actions": [
        "browser element actions"
      ],
      "counts": [
        {
          "basis": "reported",
          "evidence_url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md",
          "population": "full VWA suite",
          "qualifier": "exact",
          "unit": "tasks",
          "value": 910
        },
        {
          "basis": "reported",
          "evidence_url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md",
          "population": "one task from each template type",
          "qualifier": "exact",
          "unit": "human trajectories",
          "value": 233
        }
      ],
      "evaluation": "Execution-based completion checks for visually grounded web tasks.",
      "evidence_urls": [
        "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md"
      ],
      "focus": [
        "multimodal web task completion"
      ],
      "notes": "The 233 demonstration trajectories are not the full 910-task evaluation set. Observation modes and captioning components differ across baseline agents.",
      "observations": [
        "screenshots with set-of-mark annotations",
        "caption-augmented accessibility tree"
      ],
      "setting": "interactive"
    }
  },
  "environments": [
    "web"
  ],
  "evidence": [
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-16T06:35:57Z",
      "role": "primary",
      "sha256": "182529c2fce1e41f3a2162727205c54da8b3d2381b2273d726170ee9d8cf65cc",
      "supports": [
        "name",
        "kind",
        "summary",
        "environments",
        "topics",
        "scope",
        "relations",
        "details.benchmark_protocol"
      ],
      "url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md"
    }
  ],
  "id": "github:web-arena-x/visualwebarena",
  "identifiers": {},
  "kind": "benchmark",
  "last_verified_at": "2026-10-16T06:35:57Z",
  "licenses": {
    "code": null,
    "data": null,
    "document": null,
    "weights": null
  },
  "links": [
    {
      "rel": "primary",
      "url": "https://github.com/web-arena-x/visualwebarena"
    }
  ],
  "name": "VisualWebArena",
  "notes": "",
  "published_at": null,
  "relations": [
    {
      "evidence_url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md",
      "target_id": "arxiv:2401.13649",
      "type": "documented_by"
    }
  ],
  "review_status": "curated",
  "scope": "core",
  "summary": "Visually grounded web tasks extending the WebArena evaluation setting.",
  "topics": [
    "evaluation",
    "screen-understanding"
  ],
  "updated_at": null,
  "url": "https://github.com/web-arena-x/visualwebarena"
}