{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.13040", "kind": "paper", "url": "https://arxiv.org/abs/2606.13040"}, {"kind": "website", "url": "https://processbench-2026.github.io/RoboProcessBench-Web/"}], "authors": ["Dayu Xia", "Yue Shi", "Yao Mu", "Huiting Ji", "Chaofan Ma", "Yingjie Zhou", "Jiezhang Cao", "Hua Chen", "Yang Liu", "Guangtao Zhai"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.13040", "source_url": "https://arxiv.org/abs/2606.13040"}, "categories": ["Multimodal", "Robot manipulation", "Robotics", "Robotics & Embodied AI"], "description": {"en": "RoboProcessBench evaluates vision-language models on process-aware understanding in robotic manipulation, with 12 diagnostic question families covering static monitoring and dynamic reasoning over execution traces. The benchmark includes 58k QA pairs across 260 tasks."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.13040", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.13040", "title": "RoboProcessBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Shanghai Artificial Intelligence Laboratory", "Zhejiang University", "Shanghai Jiao Tong University", "China University of Mining and Technology", "Tsinghua University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.13040", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.13040", "source_url": "https://openalex.org/W7164577801"}, "key": "claire-radar:2606.13040", "modality": null, "name": "RoboProcessBench", "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_roboprocessbench_629af272", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "677465831cd58980bc79753bc31423ce1341c6cab0f84cead07ea61419fbdcbd", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.13040"}, "publisher": null, "released": "2026-06-11", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.13040", "source_url": "https://arxiv.org/abs/2606.13040"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-13040", "source": "claire_radar", "source_benchmark_id": "2606.13040", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Robotics & Embodied AI"], "area": "Robotics & Embodied AI", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 0, "hfPaperUrl": "https://huggingface.co/papers/2606.13040", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-10-01", "state": "stale"}}, "status": "stale"}, "capabilities": ["Robot manipulation"], "confidence": "High", "construction": "Transform Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "9967320f876bc6201dd5f2694cb322a41863de83122a184d7a74dcd26496a1e0", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "RoboProcessBench evaluates vision-language models on process-aware understanding in robotic manipulation, with 12 diagnostic question families covering static monitoring and dynamic reasoning over execution traces. The benchmark includes 58k QA pairs across 260 tasks.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "To address this gap, we present RoboProcessBench, a benchmark for process-aware understanding in vision-language robotic manipulation."}, "familyId": "bmf_51d8097c561d", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_roboprocessbench_629af272", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Robotics"], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2606.13040", "paper": "https://arxiv.org/abs/2606.13040", "pdf": "https://arxiv.org/pdf/2606.13040", "project": "https://processbench-2026.github.io/RoboProcessBench-Web/", "report": "https://arxiv.org/abs/2606.13040"}, "metrics": [], "motivation": "Vision-language models (VLMs) are increasingly explored as visual critics, reward generators, and failure detectors in robotic manipulation.", "name": "RoboProcessBench", "oneLine": "RoboProcessBench evaluates vision-language models on process-aware understanding in robotic manipulation, with 12 diagnostic question families covering static monitoring and dynamic reasoning over execution traces. The benchmark includes 58k QA pairs across 260 tasks.", "paperTitle": "RoboProcessBench: Benchmarking Process-Aware Understanding in Vision-Language Robotic Manipulation", "primaryDomain": "Robotics & Embodied AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.13040", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "RoboProcessBench Project", "organizationType": "community", "role": "benchmark-publisher", "sourceUrl": "https://processbench-2026.github.io/RoboProcessBench-Web/"}], "ranking": {}, "readiness": "Inspectable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-06-11", "source": {"authors": ["Dayu Xia", "Yue Shi", "Yao Mu", "Huiting Ji", "Chaofan Ma", "Yingjie Zhou", "Hua Chen", "Yang Liu", "Jiezhang Cao", "Guangtao Zhai"], "categories": ["cs.RO"], "id": "2606.13040", "title": "RoboProcessBench: Benchmarking Process-Aware Understanding in Vision-Language Robotic Manipulation", "type": "arxiv", "url": "https://arxiv.org/abs/2606.13040"}, "sourceUpdatedAt": "2026-08-04T14:27:50Z", "topics": ["Robotics", "Multimodal"], "venueAttempts": [], "whyItMatters": "Existing evaluations largely ignore fine-grained process understanding, which is crucial for VLMs used as critics or failure detectors. This benchmark provides a structured way to measure progress in this capability and supports post-training via a dedicated SFT split."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
