{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2605.28579", "kind": "paper", "url": "https://arxiv.org/abs/2605.28579"}, {"kind": "website", "url": "https://dong7313.github.io/muse-benchmark/"}], "authors": ["Xiaoyu Dong", "Zeng Chenbo", "Zhi Li", "Xiao-Ming Wu"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2605.28579", "source_url": "https://arxiv.org/abs/2605.28579"}, "categories": ["CAD", "General AI", "Geometric reasoning", "Science & Engineering"], "description": {"en": "MUSE evaluates text-to-CAD generation of complex B-Rep assemblies via design specifications. It scores models on code validity, geometric correctness, and design-intent alignment using rubrics covering functionality, manufacturability, and assemblability. A VLM judge with human validation is used for scalable scoring."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2605.28579", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2605.28579", "title": "MUSE"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2605.28579", "modality": null, "name": "MUSE", "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_muse_8c1c770b", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "602b1083d3d1e6434ccdb7bfbdf420a6de987140458d74146d13ddf5ce03dcb8", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2605.28579"}, "publisher": null, "released": "2026-05-27", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2605.28579", "source_url": "https://arxiv.org/abs/2605.28579"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2605-28579", "source": "claire_radar", "source_benchmark_id": "2605.28579", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Science & Engineering", "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/2605.28579", "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": "2026-09-30", "state": "stale"}}, "status": "stale"}, "capabilities": ["Geometric reasoning"], "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": "0397b51cc33fa03c4e4e04f9ab59cf3a3d9d9e5a19cbcf53a731bb6a569cb245", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "MUSE evaluates text-to-CAD generation of complex B-Rep assemblies via design specifications. It scores models on code validity, geometric correctness, and design-intent alignment using rubrics covering functionality, manufacturability, and assemblability. A VLM judge with human validation is used for scalable scoring.", "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", "public artifact URL"], "snippet": "To address this gap, we introduce MUSE, a Text-to-CAD benchmark focused on complex, editable boundary representation (B-Rep) assemblies."}, "familyId": "bmf_4016c3db3bc3", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 2, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "candidates"}, "heat": null, "id": "bm_muse_8c1c770b", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2605.28579", "paper": "https://arxiv.org/abs/2605.28579", "pdf": "https://arxiv.org/pdf/2605.28579", "project": "https://dong7313.github.io/muse-benchmark/", "report": "https://arxiv.org/abs/2605.28579"}, "metrics": [], "motivation": "Large language models (LLMs) have recently advanced text-driven 3D generation, yet Text-to-CAD remains far from supporting industrial product design.", "name": "MUSE", "oneLine": "MUSE evaluates text-to-CAD generation of complex B-Rep assemblies via design specifications. It scores models on code validity, geometric correctness, and design-intent alignment using rubrics covering functionality, manufacturability, and assemblability. A VLM judge with human validation is used for scalable scoring.", "paperTitle": "MUSE: Benchmarking Manufacturable, Functional, and Assemblable Text-to-CAD Generation", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2605.28579", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "MUSE Benchmark Team", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://dong7313.github.io/muse-benchmark/"}], "ranking": {}, "readiness": "Inspectable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-05-27", "source": {"authors": ["Xiaoyu Dong", "Zhi Li", "Xiao-Ming Wu"], "categories": ["cs.AI"], "id": "2605.28579", "title": "MUSE: Benchmarking Manufacturable, Functional, and Assemblable Text-to-CAD Generation", "type": "arxiv", "url": "https://arxiv.org/abs/2605.28579"}, "sourceUpdatedAt": "2026-06-04T01:52:56Z", "topics": ["CAD"], "venueAttempts": [], "whyItMatters": "Existing CAD benchmarks focus on single-part geometric similarity, missing industrial requirements. MUSE provides a structured evaluation that measures practical design quality, enabling progress toward engineering-ready CAD generation."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
