{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.00188", "kind": "paper", "url": "https://arxiv.org/abs/2606.00188"}, {"id": "gh:paintbench/paintbench", "kind": "repo", "url": "https://github.com/PaintBench/PaintBench"}, {"kind": "website", "url": "https://paintbench.github.io/"}], "authors": ["Kai Xu", "Ellis Brown", "Shrikar Madhu", "Rob Fergus", "He He", "Saining Xie"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.00188", "source_url": "https://arxiv.org/abs/2606.00188"}, "categories": ["Reasoning", "Robot manipulation", "Robotics", "Robotics & Embodied AI"], "description": {"en": "PaintBench is a procedurally generated benchmark for precise visual editing, covering 20 tasks across geometric, structural, color, and symbolic categories. Stable scoring uses pixel-level mIoU with fixed seeds."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.00188", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.00188", "title": "PaintBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["New York University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.00188", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.00188", "source_url": "https://openalex.org/W7163156560"}, "key": "claire-radar:2606.00188", "modality": null, "name": "PaintBench", "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_paintbench_3ad5e836", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "4f8115a1215e38b0f93231c5080b07a185df9e82a2b71d864f0b55389cb28bc7", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.00188"}, "publisher": null, "released": "2026-05-29", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.00188", "source_url": "https://arxiv.org/abs/2606.00188"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-00188", "source": "claire_radar", "source_benchmark_id": "2606.00188", "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": "https://github.com/PaintBench/PaintBench", "githubScope": "benchmark_repo", "githubStars": 5, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 4, "hfPaperUrl": "https://huggingface.co/papers/2606.00188", "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": "2026-09-30", "state": "stale"}}, "status": "fresh"}, "capabilities": ["Reasoning", "Robot manipulation"], "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": "3b9ce3e72b51dc69a158da08902c51bd2b3e1556ac4ed7da93c39ad4297e7cd3", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "PaintBench is a procedurally generated benchmark for precise visual editing, covering 20 tasks across geometric, structural, color, and symbolic categories. Stable scoring uses pixel-level mIoU with fixed seeds.", "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 probe this challenge, we introduce PaintBench, a dynamically scalable benchmark targeting 20 fundamental precise visual editing operations across four categories: geometric transformation, structural manipulation, color change, and symbolic reasoning."}, "familyId": "bmf_3505fc50bfea", "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/PaintBench/PaintBench", "url": "https://github.com/PaintBench/PaintBench"}, {"basis": "existing-metric-association", "linkedUrl": "https://github.com/PaintBench/PaintBench", "url": "https://huggingface.co/papers/2606.00188"}, {"basis": "linked-from-project", "linkedUrl": "https://github.com/PaintBench/PaintBench", "url": "https://paintbench.github.io/"}], "scope": "benchmark_repo", "url": "https://github.com/PaintBench/PaintBench", "verification": "source-linked"}], "status": "linked"}, "heat": null, "id": "bm_paintbench_3ad5e836", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Robotics"], "links": {"code": "https://github.com/PaintBench/PaintBench", "data": null, "hfPaper": "https://huggingface.co/papers/2606.00188", "paper": "https://arxiv.org/abs/2606.00188", "pdf": "https://arxiv.org/pdf/2606.00188", "project": "https://paintbench.github.io/", "report": "https://arxiv.org/abs/2606.00188"}, "metrics": [], "motivation": "While current multimodal models are proficient at open-ended visual editing, executing precise single-answer edits remains an important obstacle.", "name": "PaintBench", "oneLine": "PaintBench is a procedurally generated benchmark for precise visual editing, covering 20 tasks across geometric, structural, color, and symbolic categories. Stable scoring uses pixel-level mIoU with fixed seeds.", "paperTitle": "PaintBench: Deterministic Evaluation of Precise Visual Editing", "primaryDomain": "Robotics & Embodied AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.00188", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "NYU", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://github.com/PaintBench/PaintBench"}], "ranking": {}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-05-29", "source": {"authors": ["Kai Xu", "Ellis Brown", "Shrikar Madhu", "Rob Fergus", "He He", "Saining Xie"], "categories": ["cs.GR", "cs.CV", "cs.LG"], "id": "2606.00188", "title": "PaintBench: Deterministic Evaluation of Precise Visual Editing", "type": "arxiv", "url": "https://arxiv.org/abs/2606.00188"}, "sourceUpdatedAt": "2026-05-29T16:01:14Z", "topics": ["Robotics", "Reasoning"], "venueAttempts": [], "whyItMatters": "Procedural generation allows contamination-resistant evaluation of precise editing capabilities, with deterministic scoring that avoids human or LLM bias."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
