{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.18239", "kind": "paper", "url": "https://arxiv.org/abs/2606.18239"}, {"id": "gh:internrobotics/ebench", "kind": "repo", "url": "https://github.com/InternRobotics/EBench"}], "authors": ["Ning Gao", "Jinliang Zheng", "Xing Gao", "Haoxiang Ma", "Hanqing Wang", "Yukai Wang", "Jiantong Chen", "Zanxin Chen", "Shujie Zhang", "Mingda Jia", "Xuekun Jiang", "Zihou Zhu", "Xinyu Li", "Shuai Wang", "Hao Li", "Wenzhe Cai", "Yuqiang Yang", "Xudong Xu", "Zhaoyang Lyu", "Yao Mu", "Tai Wang", "Jiangmiao Pang", "Jia Zeng", "Weinan Zhang", "Chunhua Shen"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.18239", "source_url": "https://arxiv.org/abs/2606.18239"}, "categories": ["Robot manipulation", "Robotics", "Robotics & Embodied AI"], "description": {"en": "EBench is a simulation benchmark for diagnosing generalist mobile manipulation policies. It comprises 26 manipulation tasks annotated along five capability dimensions (scene, atomic skill, horizon, precision, mobility) and four generalization dimensions (object, background, instruction, mixed), with strict train/test splits and held-out online evaluation."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.18239", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.18239", "title": "EBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Shanghai Artificial Intelligence Laboratory", "Xi'an Jiaotong University", "Tsinghua University", "University of Science and Technology of China", "Shanghai Jiao Tong University", "Zhejiang University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.18239", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.18239", "source_url": "https://openalex.org/W7165043599"}, "key": "claire-radar:2606.18239", "modality": null, "name": "EBench", "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_ebench_35561bef", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "73c1a094281b4be4f312213dea2ce6e55499d017dcb56f36d212f6d2c660d9e1", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.18239"}, "publisher": null, "released": "2026-06-16", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.18239", "source_url": "https://arxiv.org/abs/2606.18239"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-18239", "source": "claire_radar", "source_benchmark_id": "2606.18239", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Robotics & Embodied AI"], "area": "Robotics & Embodied AI", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/InternRobotics/EBench", "githubScope": "benchmark_repo", "githubStars": 145, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 17, "hfPaperUrl": "https://huggingface.co/papers/2606.18239", "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": "2026-10-02", "state": "stale"}}, "status": "fresh"}, "capabilities": ["Robot manipulation"], "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": "2b36e4ea16e26ac2f96a4b76eeeb920cb7cc763be748ef428da62af36d77d8c5", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "EBench is a simulation benchmark for diagnosing generalist mobile manipulation policies. It comprises 26 manipulation tasks annotated along five capability dimensions (scene, atomic skill, horizon, precision, mobility) and four generalization dimensions (object, background, instruction, mixed), with strict train/test splits and held-out online evaluation.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We present EBench, a simulation benchmark that diagnoses generalist mobile manipulation policies beyond a single success-rate scalar."}, "familyId": "bmf_e19c9d617150", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [{"evidence": [{"basis": "existing-code-link", "linkedUrl": "https://github.com/InternRobotics/EBench", "url": "https://github.com/InternRobotics/EBench"}, {"basis": "existing-metric-association", "linkedUrl": "https://github.com/InternRobotics/EBench", "url": "https://huggingface.co/papers/2606.18239"}, {"basis": "huggingface-paper-association", "linkedUrl": "https://github.com/InternRobotics/EBench", "url": "https://huggingface.co/api/papers/2606.18239"}], "scope": "benchmark_repo", "url": "https://github.com/InternRobotics/EBench", "verification": "source-linked"}], "status": "linked"}, "heat": null, "id": "bm_ebench_35561bef", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Robotics"], "links": {"code": "https://github.com/InternRobotics/EBench", "data": null, "hfPaper": "https://huggingface.co/papers/2606.18239", "paper": "https://arxiv.org/abs/2606.18239", "pdf": "https://arxiv.org/pdf/2606.18239", "project": null, "report": "https://arxiv.org/abs/2606.18239"}, "metrics": [], "motivation": "We present EBench, a simulation benchmark that diagnoses generalist mobile manipulation policies beyond a single success-rate scalar.", "name": "EBench", "oneLine": "EBench is a simulation benchmark for diagnosing generalist mobile manipulation policies. It comprises 26 manipulation tasks annotated along five capability dimensions (scene, atomic skill, horizon, precision, mobility) and four generalization dimensions (object, background, instruction, mixed), with strict train/test splits and held-out online evaluation.", "paperTitle": "EBench: Elemental Diagnosis of Generalist Mobile Manipulation Policies", "primaryDomain": "Robotics & Embodied AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.18239", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "Shanghai AI Laboratory", "organizationType": "company-research-lab", "role": "benchmark-publisher", "sourceUrl": "https://github.com/InternRobotics/EBench"}], "ranking": {}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-06-16", "source": {"authors": ["Ning Gao", "Jinliang Zheng", "Xing Gao", "Haoxiang Ma", "Hanqing Wang", "Yukai Wang", "Jiantong Chen", "Zanxin Chen", "Shujie Zhang", "Mingda Jia", "Xuekun Jiang", "Zihou Zhu", "Xinyu Li", "Shuai Wang", "Hao Li", "Wenzhe Cai", "Yuqiang Yang", "Xudong Xu", "Zhaoyang Lyu", "Yao Mu", "Tai Wang", "Jiangmiao Pang", "Jia Zeng", "Weinan Zhang", "Chunhua Shen"], "categories": ["cs.RO"], "id": "2606.18239", "title": "EBench: Elemental Diagnosis of Generalist Mobile Manipulation Policies", "type": "arxiv", "url": "https://arxiv.org/abs/2606.18239"}, "sourceUpdatedAt": "2026-06-20T06:00:15Z", "topics": ["Robotics"], "venueAttempts": [], "whyItMatters": "EBench addresses the need for multi-axis diagnostic evaluation of generalist manipulation models, moving beyond single success-rate metrics. It provides practical decision value by revealing capability and generalization profiles that guide model iteration and selection."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
