{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.23329", "kind": "paper", "url": "https://arxiv.org/abs/2608.23329v1"}], "authors": ["Wenqi Liu", "Shijie Ma", "Yunxiao Wang", "Meng Liu", "Qile Su", "Han Liu", "Bohan Hou", "Zeyu Wang", "Xuanyu Zheng", "Changyi Liu", "Tianke Zhang", "Haonan Fan", "Kaiyu Jiang", "Yingxin Li", "Jiankang Chen", "Xu Wang", "Hongyi Fu", "Jianxiong Wang", "Bin Wen", "Tingting Gao", "Han Li", "Jianhua Yin", "Yinwei Wei", "Xuemeng Song"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.23329", "source_url": "https://arxiv.org/abs/2608.23329"}, "categories": ["General AI", "Reasoning", "Vision & 3D"], "description": {"en": "VideoRover-Bench is a benchmark for open-world video reasoning that combines video understanding with deep research, stratified by video duration and research difficulty. It evaluates the capability of models to locate sparse visual evidence and acquire external knowledge through coordinated video cropping, multimodal search, and webpage browsing."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.23329v1", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.23329v1", "title": "VideoRover-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Shandong University", "Chinese Academy of Sciences", "Institute of Automation", "Beihang University", "City University of Hong Kong", "Nanyang Technological University", "Kuaishou (China)", "Southern University of Science and Technology"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.23329", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.23329", "source_url": "https://openalex.org/W7204203119"}, "key": "claire-radar:2608.23329", "modality": null, "name": "VideoRover-Bench", "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_videorover-bench_f87e9ddf", "origin_source": "arxiv", "recognition_confidence": "0.65", "record_sha256": "f08656c514feb2c0cd2c5772c262f65be9bb8c7281d3da4bb257c773a1ab5a87", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-08-27T04:18:56.318436Z", "source_url": "https://arxiv.org/abs/2608.23329v1"}, "publisher": null, "released": "2026-08-24", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.23329", "source_url": "https://arxiv.org/abs/2608.23329v1"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-23329", "source": "claire_radar", "source_benchmark_id": "2608.23329", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Vision & 3D", "attention": {"asOf": "2026-08-26", "attemptedAt": "2026-08-25T23:36:26.005677Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2608.23329", "observedAt": "2026-08-24T11:31:30.653116Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-08-25T23:36:26.005677Z", "state": "unavailable"}}, "status": "unavailable"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "The benchmark addresses trending topics in video understanding and agentic AI, with clear release and evaluation framework.", "score": 55}, "capabilities": ["Reasoning"], "confidence": "Medium", "construction": "Transform Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-27T04:22:25.189955Z", "inputHash": "b945bd507dc1d7358272622c9e1fb672358f5aa810b507c2a1892c63f8da16aa", "model": "deepseek-v4-flash"}, "curation": {"canonicalNameEvidence": "We also introduce VideoRover-Bench, a benchmark stratified by video duration and research difficulty.", "canonicalNameSource": "abstract", "decisionReason": "Formally named benchmark with public evaluation protocol used to compare models and tool configurations.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-27T04:18:56.318436Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "VideoRover-Bench is a benchmark for open-world video reasoning that combines video understanding with deep research, stratified by video duration and research difficulty. It evaluates the capability of models to locate sparse visual evidence and acquire external knowledge through coordinated video cropping, multimodal search, and webpage browsing.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "evaluation protocol evidence"], "snippet": "We also introduce VideoRover-Bench, a benchmark stratified by video duration and research difficulty."}, "familyId": "bmf_f0af8d10d7cf", "firstSeenAt": "2026-08-25", "heat": null, "id": "bm_videorover-bench_f87e9ddf", "indexedAt": "2026-08-25T16:21:05.134052Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "http://arxiv.org/abs/2608.23329v1", "pdf": "https://arxiv.org/pdf/2608.23329v1", "project": null, "report": "http://arxiv.org/abs/2608.23329v1"}, "metrics": [], "motivation": "Open-world video understanding often requires a model to locate sparse visual evidence and acquire external knowledge that is absent from the video and its parametric memory.", "name": "VideoRover-Bench", "oneLine": "VideoRover-Bench is a benchmark for open-world video reasoning that combines video understanding with deep research, stratified by video duration and research difficulty. It evaluates the capability of models to locate sparse visual evidence and acquire external knowledge through coordinated video cropping, multimodal search, and webpage browsing.", "paperTitle": "Thinking Beyond Videos: Unifying Video Reasoning and Deep Research for Open-World Video Agents", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "http://arxiv.org/abs/2608.23329v1", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-25T16:21:05.134052Z"}, "publications": [], "publishers": [{"name": "VideoRover Team", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "http://arxiv.org/abs/2608.23329v1"}], "ranking": {"30d": {"confidence": "Low", "coverage": 0.0, "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": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "rank": null, "score": null}, "method": "attention-ranking-v11", "rank": null, "score": null}, "90d": {"confidence": "Low", "coverage": 0.0, "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": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "rank": null, "score": null}, "method": "attention-ranking-v11", "rank": null, "score": null}}, "readiness": "Paper only", "recognitionConfidence": 0.65, "relation": "introduces", "releasedAt": "2026-08-24", "source": {"authors": ["Wenqi Liu", "Shijie Ma", "Yunxiao Wang", "Meng Liu", "Qile Su", "Han Liu", "Bohan Hou", "Xuanyu Zheng", "Changyi Liu", "Tianke Zhang", "Haonan Fan", "Kaiyu Jiang", "Yingxin Li", "Jiankang Chen", "Xu Wang", "Bin Wen", "Tingting Gao", "Han Li", "Jianhua Yin", "Yinwei Wei", "Xuemeng Song"], "categories": ["cs.CV", "cs.AI"], "id": "2608.23329", "title": "Thinking Beyond Videos: Unifying Video Reasoning and Deep Research for Open-World Video Agents", "type": "arxiv", "url": "http://arxiv.org/abs/2608.23329v1"}, "sourceUpdatedAt": "2026-08-24T14:42:23Z", "topics": ["Reasoning"], "venueAttempts": [], "whyItMatters": "The benchmark fills a gap by assessing unified video reasoning and multi-step information seeking, which are typically developed in isolation. It provides a structured evaluation to guide the development of video agents that can handle complex open-world tasks requiring external knowledge."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
