{"record": {"aliases": ["Enginuity: A Dataset and Benchmark for Vision-Language Understanding of Engineering Diagrams"], "artifacts": [{"id": "arxiv:2606.03410", "kind": "paper", "url": "https://arxiv.org/abs/2606.03410"}], "categories": ["General AI", "Multimodal"], "description": {"en": "Enginuity evaluates vision-language models on structured parts-table extraction and visual question answering over complex engineering diagrams."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.03410", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.03410", "title": "Enginuity"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2606.03410", "modality": null, "name": "Enginuity", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_enginuity-a-dataset-and-benchmark-for-visi_a1a985ac", "origin_source": "arxiv", "recognition_confidence": "0.55", "record_sha256": "5c78ef47fb0197432b65a91c055fa71ca9e019650fd8b00733d0bf77b343c08f", "relation": "evaluates_only", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T19:58:48.870689Z", "source_url": "https://arxiv.org/abs/2606.03410"}, "publisher": null, "released": "2026-06-02", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.03410", "source_url": "https://arxiv.org/abs/2606.03410"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-03410", "source": "claire_radar", "source_benchmark_id": "2606.03410", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["Enginuity: A Dataset and Benchmark for Vision-Language Understanding of Engineering Diagrams"], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Multimodal", "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "The specialized engineering-diagram focus may attract vision-language evaluation researchers, though its near-term audience is likely narrower than that of general-purpose multimodal benchmarks.", "score": 4}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T19:58:48.870689Z", "inputHash": "d3f1e22e8a8d1baf271933974a62599a3e7df6dad1314dbe01eb1613cc6ab994", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "Enginuity", "canonicalNameSource": "paper_title", "decisionReason": "The source formally names the benchmark, defines two repeatable tasks, reports concrete comparable metrics, and states that the dataset, annotations, evaluation harness, and model outputs are released.", "evaluationEvidence": {"reuse": "We release the dataset, annotations, evaluation harness, and per-sample model outputs", "scoring": "Recall@all of 0.61-0.87 but Token F1pen of only 0.03-0.18", "task": "structured parts-table extraction (Task 1) and free-form visual diagram question answering (VQA)(Task 2)"}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T19:58:48.870689Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Enginuity evaluates vision-language models on structured parts-table extraction and visual question answering over complex engineering diagrams.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence", "named entity is a model, method, framework, or dataset used in evaluation"], "snippet": "To address this gap, we introduce Enginuity, the first open dataset and benchmark for evaluating VLMs on complex engineering diagrams."}, "familyId": "bmf_50819ec0f97f", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_enginuity-a-dataset-and-benchmark-for-visi_a1a985ac", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2606.03410", "pdf": "https://arxiv.org/pdf/2606.03410", "project": null, "report": "https://arxiv.org/abs/2606.03410"}, "metrics": [], "motivation": "Engineering diagrams pose a distinct challenge for vision-language models: unlike natural images or general documents, they encode information through dense spatial layouts, domain-specific symbols, and cross-references between visual callouts and structured parts tables.", "name": "Enginuity", "oneLine": "Enginuity evaluates vision-language models on structured parts-table extraction and visual question answering over complex engineering diagrams.", "paperTitle": "Enginuity: A Dataset and Benchmark for Vision-Language Understanding of Engineering Diagrams", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.03410", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "readiness": "Paper only", "recognitionConfidence": 0.55, "relation": "evaluates_only", "releasedAt": "2026-06-02", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Abhishek Kumar", "Isha Motiyani", "Tilak Kasturi", "Ethan Seefried", "Prahitha Movva", "Tirthankar Ghosal"], "categories": ["cs.CV"], "id": "2606.03410", "title": "Enginuity: A Dataset and Benchmark for Vision-Language Understanding of Engineering Diagrams", "type": "arxiv", "url": "https://arxiv.org/abs/2606.03410"}, "sourceUpdatedAt": "2026-06-02T09:54:03Z", "topics": ["Multimodal"], "venueAttempts": [], "whyItMatters": "It provides a reproducible evaluation of model capabilities on dense technical diagrams with domain-specific symbols, spatial layouts, and parts-table references."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
