{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.29537", "kind": "paper", "url": "https://arxiv.org/abs/2606.29537"}, {"id": "gh:xlang-ai/osworld-v2", "kind": "repo", "url": "https://github.com/xlang-ai/OSWorld-V2"}], "authors": ["Mengqi Yuan", "Zilong Zhou", "Xinzhuang Xiong", "Weiming Wu", "Jiayang Sun", "Jiamin Song", "Kaiqian Cui", "Bowen Wang", "Haoyuan Wu", "Yitong Li", "Dunjie Lu", "Haikong Lu", "Qi Zhen", "Xinyuan Wang", "Jiaqi Deng", "Yuhao Yang", "Cheng Chen", "Boyuan Zheng", "Alex Su", "Xiao Yu", "Hao Zou", "Saaket Agashe", "Xing Han Lu", "Manpreet Kaur", "Zhengyang Qi", "Vincent Sunn Chen", "Frederic Sala", "Dayiheng Liu", "Junyang Lin", "Zhou Yu", "Yu Su", "Siva Reddy", "Xin Eric Wang", "Peng Qi", "Tianbao Xie", "Tao Yu"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.29537", "source_url": "https://arxiv.org/abs/2606.29537"}, "categories": ["Agents & Tool Use", "General AI", "cs.AI"], "description": {"en": "Evaluates computer-use agents on 108 long-horizon real-world workflows across everyday and professional tasks, scored by binary completion at 500 steps and partial scores."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.29537", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.29537", "title": "OSWorld 2.0"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2606.29537", "modality": null, "name": "OSWorld 2.0", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_osworld-2-0_ee079e02", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "d678ce837bfafb5bcca83bd0ffc409efe36e16d0eae54553e9604895dc740263", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.29537"}, "publisher": null, "released": "2026-06-28", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.29537", "source_url": "https://arxiv.org/abs/2606.29537"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-29537", "source": "claire_radar", "source_benchmark_id": "2606.29537", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Agents & Tool Use", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/xlang-ai/OSWorld-V2", "githubScope": "benchmark_repo", "githubStars": 350, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 24, "hfPaperUrl": "https://huggingface.co/papers/2606.29537", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-10-04", "state": "stale"}}, "status": "fresh"}, "capabilities": [], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "6a46fc7ce1c8ae35070e7444ec1e6176ab91d714e98416ae04dc584ef408e356", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates computer-use agents on 108 long-horizon real-world workflows across everyday and professional tasks, scored by binary completion at 500 steps and partial scores.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact coined title identity tied to benchmark evidence", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We introduce OSWorld 2.0, a benchmark of 108 long-horizon computer-use workflows across everyday and professional tasks, designed to capture complex and challenging real-world phenomena."}, "familyId": "bmf_f1ccc6d9d057", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [{"evidence": [{"basis": "existing-code-link", "linkedUrl": "https://github.com/xlang-ai/OSWorld-V2", "url": "https://github.com/xlang-ai/OSWorld-V2"}, {"basis": "existing-metric-association", "linkedUrl": "https://github.com/xlang-ai/OSWorld-V2", "url": "https://huggingface.co/papers/2606.29537"}], "scope": "project_repo", "url": "https://github.com/xlang-ai/OSWorld-V2", "verification": "source-linked"}], "status": "linked"}, "heat": null, "id": "bm_osworld-2-0_ee079e02", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/xlang-ai/OSWorld-V2", "data": null, "hfPaper": "https://huggingface.co/papers/2606.29537", "paper": "https://arxiv.org/abs/2606.29537", "pdf": "https://arxiv.org/pdf/2606.29537", "project": null, "report": "https://arxiv.org/abs/2606.29537"}, "metrics": [], "motivation": "Existing computer-use benchmarks fail to capture the realism, complexity, and long-horizon demands of real-world computer use, limiting their ability to reveal the limitations of frontier agents.", "name": "OSWorld 2.0", "oneLine": "Evaluates computer-use agents on 108 long-horizon real-world workflows across everyday and professional tasks, scored by binary completion at 500 steps and partial scores.", "paperTitle": "OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.29537", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "XLang Lab", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://github.com/xlang-ai/OSWorld-V2"}], "ranking": {}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-06-28", "source": {"authors": ["Mengqi Yuan", "Zilong Zhou", "Xinzhuang Xiong", "Weiming Wu", "Jiayang Sun", "Jiamin Song", "Kaiqian Cui", "Bowen Wang", "Haoyuan Wu", "Yitong Li", "Dunjie Lu", "Haikong Lu", "Qi Zhen", "Xinyuan Wang", "Jiaqi Deng", "Yuhao Yang", "Cheng Chen", "Boyuan Zheng", "Alex Su", "Xiao Yu", "Hao Zou", "Saaket Agashe", "Xing Han Lu", "Manpreet Kaur", "Zhengyang Qi", "Vincent Sunn Chen", "Frederic Sala", "Dayiheng Liu", "Junyang Lin", "Zhou Yu", "Yu Su", "Siva Reddy", "Xin Eric Wang", "Peng Qi", "Tianbao Xie", "Tao Yu"], "categories": ["cs.AI"], "id": "2606.29537", "title": "OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks", "type": "arxiv", "url": "https://arxiv.org/abs/2606.29537"}, "sourceUpdatedAt": "2026-07-13T10:30:17Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "Captures long-horizon, dynamic, and hidden-state challenges absent in prior benchmarks, revealing that agents fail on constraint tracking and mid-task information."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
