{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.19583", "kind": "paper", "url": "https://arxiv.org/abs/2608.19583"}], "authors": ["Xuan He", "Cong Wei", "Yuhao Cheng", "Linrui Ma", "Yuxuan Zhang", "Zuojun Li", "Yuhao Wen", "Jize Jiang", "Zeyi Liu", "Yuren Hao", "Songcheng Cai", "Keming Wu", "Penghui Du", "Kai Zou", "Rui Yang", "Chenkai Sun", "Ke Yang", "Ping Nie", "Kelsey R Allen", "Chenglong Wang", "Michel Galley", "Jianfeng Gao", "ChengXiang Zhai"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.19583", "source_url": "https://arxiv.org/abs/2608.19583"}, "categories": ["General AI", "Reasoning", "Vision & 3D"], "description": {"en": "Contains 27 tasks and 810 instances organized by task domains and skill tags to evaluate visual reasoning capabilities of video generation models."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.19583", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.19583", "title": "VGI-BENCH"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.19583", "modality": null, "name": "VGI-BENCH", "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": "true", "origin_record_id": "bm_vgi-bench_4a5d68ac", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "39adebb1d34010aa9c8ddb4ccaf832513d013528bb3018667103485038a2164d", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-08-25T17:43:56.718736Z", "source_url": "https://arxiv.org/abs/2608.19583"}, "publisher": null, "released": "2026-08-20", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.19583", "source_url": "https://arxiv.org/abs/2608.19583"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-19583", "source": "claire_radar", "source_benchmark_id": "2608.19583", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Machine Generated", "applicationDomains": ["General AI"], "area": "Vision & 3D", "attention": {"asOf": "2026-08-26", "attemptedAt": "2026-08-25T16:48:55.714492Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 0, "hfPaperUrl": "https://huggingface.co/papers/2608.19583", "observedAt": "2026-08-25T16:48:55.714492Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"observedAt": "2026-08-25T16:48:55.714492Z", "state": "fresh"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "The benchmark targets a trending area of video generation and reports strong model performance gaps, likely to attract substantial research attention.", "score": 80}, "capabilities": ["Reasoning"], "confidence": "High", "construction": "Original Synthetic", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-25T17:43:56.718736Z", "inputHash": "cc2312da59effb404c76dffabac3e318f5329779269455ecdee62b961daa4485", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "VGI-BENCH: Probing Visual Intelligence in Video Generation Models", "canonicalNameSource": "paper_title", "decisionReason": "The paper introduces a formal benchmark with a clear taxonomy and evaluation criteria, and the website provides a reuse path for other teams.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-25T17:43:56.718736Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Contains 27 tasks and 810 instances organized by task domains and skill tags to evaluate visual reasoning capabilities of video generation models.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "To this end, we introduce VGI-bench, containing 27 tasks and 810 instances, organized by a two-level taxonomy of task domains and skill tags for fine-grained evaluation of visual reasoning capabilities of video generation models."}, "familyId": "bmf_756186ceeff4", "firstSeenAt": "2026-08-21", "heat": null, "id": "bm_vgi-bench_4a5d68ac", "indexedAt": "2026-08-21T04:30:09.569171Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2608.19583", "pdf": "https://arxiv.org/pdf/2608.19583", "project": null, "report": "https://arxiv.org/abs/2608.19583"}, "metrics": [], "motivation": "Recent studies suggest that video generation models can exhibit certain forms of zero-shot visual reasoning through generated frames.", "name": "VGI-BENCH", "oneLine": "Contains 27 tasks and 810 instances organized by task domains and skill tags to evaluate visual reasoning capabilities of video generation models.", "paperTitle": "VGI-BENCH: Probing Visual Intelligence in Video Generation Models", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.19583", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-21T04:30:09.569171Z"}, "publications": [], "ranking": {"30d": {"confidence": "Low", "coverage": 0.3, "growth": {"components": {"githubStars": {"percentile": null, "value": null}, "hfDatasetDownloads": {"percentile": null, "value": null}, "hfPaperUpvotes": {"percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 30}, "level": {"components": {"githubStars": {"percentile": null, "value": null}, "hfDatasetDownloads": {"percentile": null, "value": null}, "hfPaperUpvotes": {"percentile": 0.1717171717171717, "value": 0}}, "confidence": "Low", "coverage": 0.3, "rank": null, "score": 40}, "method": "attention-ranking-v11", "rank": null, "score": 40}, "90d": {"confidence": "Low", "coverage": 0.15, "growth": {"components": {"githubStars": {"percentile": null, "value": null}, "hfDatasetDownloads": {"percentile": null, "value": null}, "hfPaperUpvotes": {"percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"percentile": null, "value": null}, "hfDatasetDownloads": {"percentile": null, "value": null}, "hfPaperUpvotes": {"percentile": 0.2119047619047619, "value": 0}}, "confidence": "Low", "coverage": 0.15, "rank": null, "score": 46}, "method": "attention-ranking-v11", "rank": null, "score": 46}}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-08-20", "source": {"authors": ["Xuan He", "Cong Wei", "Yuhao Cheng", "Linrui Ma", "Yuxuan Zhang", "Zuojun Li", "Yuhao Wen", "Zeyi Liu", "Yuren Hao", "Songcheng Cai", "Keming Wu", "Penghui Du", "Kai Zou", "Rui Yang", "Chenkai Sun", "Ke Yang", "Ping Nie", "Kelsey R Allen", "Chenglong Wang", "Michel Galley", "Jianfeng Gao", "ChengXiang Zhai"], "categories": ["cs.CV", "cs.AI"], "id": "2608.19583", "title": "VGI-BENCH: Probing Visual Intelligence in Video Generation Models", "type": "arxiv", "url": "https://arxiv.org/abs/2608.19583"}, "sourceUpdatedAt": "2026-08-20T02:56:56Z", "topics": ["Reasoning"], "venueAttempts": [], "whyItMatters": "Provides fine-grained evaluation of video generation models' visual reasoning, addressing input alignment, process validity, and task difficulty calibration."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
