{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2605.22645", "kind": "paper", "url": "https://arxiv.org/abs/2605.22645"}], "authors": ["Hanjun Luo", "Zhimu Huang", "Sylvia Chung", "Yiran Wang", "Yingbin Jin", "Jialin Li", "Jiang Li", "Xinfeng Li", "Hanan Salam"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2605.22645", "source_url": "https://arxiv.org/abs/2605.22645"}, "categories": ["General AI", "Vision & 3D", "cs.AI"], "description": {"en": "AtelierEval evaluates prompting proficiency of humans and MLLMs for text-to-image systems via 360 tasks, using a skill-based agentic evaluator (AtelierJudge) that scores prompt-image pairs subjectively and objectively."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2605.22645", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2605.22645", "title": "AtelierEval"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2605.22645", "modality": null, "name": "AtelierEval", "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_ateliereval_cc1f5b0c", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "d720c99dfa847012aedec5b65392afb537d6ac26dc14bcd6e47dde9079be8b63", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2605.22645"}, "publisher": null, "released": "2026-05-21", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2605.22645", "source_url": "https://arxiv.org/abs/2605.22645"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2605-22645", "source": "claire_radar", "source_benchmark_id": "2605.22645", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Vision & 3D", "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": null, "hfPaperUrl": "https://huggingface.co/papers/2605.22645", "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", "state": "unavailable"}}, "status": "unavailable"}, "capabilities": [], "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": "849add50d73fcb34de0a5434ed011db8056f58f81af2ee4ea6b4c4d5062dea81", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "AtelierEval evaluates prompting proficiency of humans and MLLMs for text-to-image systems via 360 tasks, using a skill-based agentic evaluator (AtelierJudge) that scores prompt-image pairs subjectively and objectively.", "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"], "snippet": "We introduce AtelierEval, the first unified benchmark that quantifies prompting proficiency across 360 expert-crafted tasks."}, "familyId": "bmf_40ece635281f", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_ateliereval_cc1f5b0c", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2605.22645", "paper": "https://arxiv.org/abs/2605.22645", "pdf": "https://arxiv.org/pdf/2605.22645", "project": null, "report": "https://arxiv.org/abs/2605.22645"}, "metrics": [], "motivation": "Text-to-image (T2I) systems increasingly rely on upstream prompters, either humans or multimodal large language models (MLLMs), to translate user intent into detailed prompts.", "name": "AtelierEval", "oneLine": "AtelierEval evaluates prompting proficiency of humans and MLLMs for text-to-image systems via 360 tasks, using a skill-based agentic evaluator (AtelierJudge) that scores prompt-image pairs subjectively and objectively.", "paperTitle": "AtelierEval: Agentic Evaluation of Humans & LLMs as Text-to-Image Prompters", "primaryDomain": "General AI", "publication": {"evidence": "Accepted by ICML 2026", "evidenceLevel": "author-claim", "evidenceUrl": "https://arxiv.org/abs/2605.22645", "source": "arxiv-comments", "status": "acceptance_claimed", "venue": "ICML 2026", "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-05-21", "source": {"authors": ["Hanjun Luo", "Zhimu Huang", "Sylvia Chung", "Yiran Wang", "Yingbin Jin", "Jialin Li", "Jiang Li", "Xinfeng Li", "Hanan Salam"], "categories": ["cs.AI"], "id": "2605.22645", "title": "AtelierEval: Agentic Evaluation of Humans & LLMs as Text-to-Image Prompters", "type": "arxiv", "url": "https://arxiv.org/abs/2605.22645"}, "sourceUpdatedAt": "2026-05-21T15:51:53Z", "topics": ["cs.AI"], "venueAttempts": [{"decisionRaw": "Accepted by ICML 2026", "evidence": [{"level": "author-claim", "observedAt": "2026-10-14T01:46:58.058199Z", "rawValue": "Accepted by ICML 2026", "sourceType": "arxiv-comments", "sourceUrl": "https://arxiv.org/abs/2605.22645"}], "reviewStatus": "accepted", "venueName": "ICML 2026"}], "whyItMatters": "Introduces a new evaluation angle for T2I pipelines, measuring upstream prompting ability which is currently unassessed."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
