{
  "aliases": [],
  "authors": [],
  "details": {
    "benchmark_protocol": {
      "actions": [
        "touch actions",
        "MCP tools (optional)"
      ],
      "counts": [
        {
          "basis": "reported",
          "evidence_url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md",
          "population": "tasks.json benchmark",
          "qualifier": "exact",
          "unit": "tasks",
          "value": 133
        },
        {
          "basis": "reported",
          "evidence_url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md",
          "population": "bundled SwiftUI apps",
          "qualifier": "exact",
          "unit": "applications",
          "value": 26
        }
      ],
      "evaluation": "Rubric satisfaction per task; full-task pass requires every criterion to be satisfied.",
      "evidence_urls": [
        "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md"
      ],
      "focus": [
        "personalized phone tasks",
        "cross-app workflows"
      ],
      "notes": "Vision-only, vision-plus-XML, and MCP modes expose different information and tools. The documented default evaluator is model-based; compare the same mode and judge.",
      "observations": [
        "screenshots",
        "accessibility XML (optional)"
      ],
      "setting": "interactive"
    },
    "evaluation_type": "rubric-based",
    "metrics": [
      "mean rubric score",
      "full-task pass rate"
    ],
    "task_count": 133
  },
  "environments": [
    "ios"
  ],
  "evidence": [
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-16T10:42:12Z",
      "role": "primary",
      "sha256": "706f9d770b65a1238e56f74f7dd5089c7756cc88e0ac09d80c7b2b3085a09b27",
      "supports": [
        "name",
        "kind",
        "summary",
        "environments",
        "topics",
        "scope",
        "relations",
        "details.task_count",
        "details.metrics",
        "details.evaluation_type",
        "details.benchmark_protocol"
      ],
      "url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md"
    },
    {
      "method": "license_header",
      "retrieved_at": "2026-10-16T10:48:41Z",
      "role": "primary",
      "sha256": "fa9fe1e31ed8da21d0b51621d3a3669ced4f0e42353aaf4bd8be1615d200eaac",
      "supports": [
        "licenses.code"
      ],
      "url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/LICENSE"
    }
  ],
  "id": "github:ljang0/iosworld",
  "identifiers": {},
  "kind": "benchmark",
  "last_verified_at": "2026-10-16T10:42:12Z",
  "licenses": {
    "code": {
      "basis": "repository LICENSE header; not a weights/data license",
      "declared": "Apache-2.0",
      "retrieved_at": "2026-10-16T10:48:41Z",
      "url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/LICENSE"
    },
    "data": null,
    "document": null,
    "weights": null
  },
  "links": [
    {
      "rel": "primary",
      "url": "https://github.com/ljang0/iOSWorld"
    }
  ],
  "name": "iOSWorld",
  "notes": "",
  "published_at": null,
  "relations": [],
  "review_status": "curated",
  "scope": "core",
  "summary": "Native iPhone-simulator benchmark with shared user data, application workflows, agent runners, and rubric-based scoring.",
  "topics": [
    "evaluation",
    "cross-app",
    "memory"
  ],
  "updated_at": null,
  "url": "https://github.com/ljang0/iOSWorld"
}