{"record": {"aliases": ["Paint What You See: Benchmarking Dexterous Visual Tool Use in Multimodal Agents"], "artifacts": [{"id": "arxiv:2608.25417", "kind": "paper", "url": "https://arxiv.org/abs/2608.25417"}], "authors": ["Shudong Liu", "Dongyang Chen", "Enci Zhang", "Jinwei Liang", "Zheng Ma", "Lewei Lu"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.25417", "source_url": "https://arxiv.org/abs/2608.25417"}, "categories": ["Agents", "Agents & Tool Use", "Multimodal", "Software & AI Compute", "Tool use"], "description": {"en": "Evaluates dexterous visual tool use through reference-guided painting, semantic annotation, handwriting, and path planning tasks."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.25417", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.25417", "title": "EASEL"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Peking University", "Sensetime (China)", "Tsinghua University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.25417", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.25417", "source_url": "https://openalex.org/W7204442068"}, "key": "claire-radar:2608.25417", "modality": null, "name": "EASEL", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_paint-what-you-see-benchmarking-dexterous-_f2ccb455", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "7d295edf7f48a4658b5a934c923e1470448c9da2f6db0223ee7ddfc517f52407", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-08-27T04:12:36.511123Z", "source_url": "https://arxiv.org/abs/2608.25417"}, "publisher": null, "released": "2026-08-26", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.25417", "source_url": "https://arxiv.org/abs/2608.25417"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-25417", "source": "claire_radar", "source_benchmark_id": "2608.25417", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["Paint What You See: Benchmarking Dexterous Visual Tool Use in Multimodal Agents"], "annotation": "Unknown", "applicationDomains": ["Software & AI Compute"], "area": "Agents & Tool Use", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/OOOHS/EASEL", "githubScope": "benchmark_repo", "githubStars": 0, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 18, "hfPaperUrl": "https://huggingface.co/papers/2608.25417", "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": "2026-09-25", "state": "stale"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "Broad multimodal agent topic with large-scale data and multiple model evaluations, though no public code link is provided.", "score": 70}, "capabilities": ["Tool use"], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-27T04:12:36.511123Z", "inputHash": "219bb05fc88581bb086f2d83b839cddcd213bffe5c0821a73a21f7e5af520ab1", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "We propose EASEL, a benchmark evaluating a controlled instance of dexterous visual tool use that adopts reference-guided visual reconstruction as its primary proxy task", "canonicalNameSource": "abstract", "decisionReason": "EASEL is explicitly proposed as a benchmark with dataset and model contributions; the paper reports evaluation of 25 models, and artifact links are absent but likely forthcoming.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-27T04:12:36.511123Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates dexterous visual tool use through reference-guided painting, semantic annotation, handwriting, and path planning tasks.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We propose EASEL, a benchmark evaluating a controlled instance of dexterous visual tool use that adopts reference-guided visual reconstruction as its primary proxy task: the agent incrementally paints a canvas to match a reference image."}, "familyId": "bmf_95ac01e427df", "firstSeenAt": "2026-08-27", "heat": null, "id": "bm_paint-what-you-see-benchmarking-dexterous-_f2ccb455", "indexedAt": "2026-08-27T04:12:10.575570Z", "industrySectors": ["Software & Cloud"], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2608.25417", "pdf": "https://arxiv.org/pdf/2608.25417", "project": null, "report": "https://arxiv.org/abs/2608.25417"}, "metrics": [], "motivation": "Evaluation is shifting from static QA toward agentic settings where models act through external tools.", "name": "EASEL", "oneLine": "Evaluates dexterous visual tool use through reference-guided painting, semantic annotation, handwriting, and path planning tasks.", "paperTitle": "Paint What You See: Benchmarking Dexterous Visual Tool Use in Multimodal Agents", "primaryDomain": "Software & AI Compute", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.25417", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-27T04:12:10.575570Z"}, "publications": [], "ranking": {"90d": {"confidence": "Medium", "coverage": 0.7, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": 0.0, "value": 0}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.5202664955757326, "value": 18}}, "confidence": "Medium", "coverage": 0.7, "rank": 385, "score": 23}, "method": "attention-ranking-v13", "rank": 385, "score": 23}}, "readiness": "Paper only", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-08-26", "source": {"authors": ["Shudong Liu", "Dongyang Chen", "Enci Zhang", "Jinwei Liang", "Zheng Ma", "Lewei Lu"], "categories": ["cs.AI"], "id": "2608.25417", "title": "Paint What You See: Benchmarking Dexterous Visual Tool Use in Multimodal Agents", "type": "arxiv", "url": "https://arxiv.org/abs/2608.25417"}, "sourceUpdatedAt": "2026-08-26T06:19:28Z", "topics": ["Agents", "Multimodal"], "venueAttempts": [], "whyItMatters": "Introduces closed-loop, parameterized visual action as an underexplored agent capability beyond static QA and navigation."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
