{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.28281", "kind": "paper", "url": "https://arxiv.org/abs/2608.28281v1"}, {"id": "gh:amap-ml/looparena", "kind": "repo", "url": "https://github.com/AMAP-ML/LoopArena"}], "authors": ["Yi Wang", "Haopeng Zhang", "Chengxiang Huang", "Rui Dai", "Kaikui Liu", "Piotr Koniusz", "Xiangxiang Chu"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.28281", "source_url": "https://arxiv.org/abs/2608.28281"}, "categories": ["Agents", "General AI", "Language & Knowledge"], "description": {"en": "Evaluates a controller model guiding a fixed coding agent through long-running tasks, scored via contract accuracy, strict success rate, and cost metrics across three complementary settings."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.28281v1", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.28281v1", "title": "LoopArena"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Alibaba Group (China)", "Commonwealth Scientific and Industrial Research Organisation", "Beijing University of Posts and Telecommunications", "UNSW Sydney", "Data61"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.28281", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.28281", "source_url": "https://openalex.org/W7204813304"}, "key": "claire-radar:2608.28281", "modality": null, "name": "LoopArena", "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_looparena_97bc3500", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "9f04bfecc4f63cb6cf6887cdc3d3417485876ec6d3786a4fa3784c19d0fa1587", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-09-05T08:28:38.611381Z", "source_url": "https://arxiv.org/abs/2608.28281v1"}, "publisher": null, "released": "2026-08-28", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.28281", "source_url": "https://arxiv.org/abs/2608.28281v1"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-28281", "source": "claire_radar", "source_benchmark_id": "2608.28281", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attention": {"asOf": "2026-09-05", "attemptedAt": "2026-09-05T01:35:52.848019Z", "githubRepo": "https://github.com/AMAP-ML/LoopArena", "githubScope": "benchmark_repo", "githubStars": 87, "hfDailySubmittedAt": "2026-08-31T00:00:00.000Z", "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 101, "hfPaperUrl": "https://huggingface.co/papers/2608.28281", "observedAt": "2026-09-05T01:35:52.848019Z", "signalStatus": {"githubStars": {"observedAt": "2026-09-05T01:35:52.848019Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"observedAt": "2026-09-05T01:35:52.848019Z", "state": "fresh"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "LoopArena targets the emerging area of coding agent orchestration with a released protocol and leaderboard, likely generating interest among LLM agent researchers and practitioners.", "score": 80}, "capabilities": [], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-05T08:28:38.611381Z", "inputHash": "33811b91cbe8e970fbda4c8cd37e946b0fbccc9798a56a7333c989aabb1cba94", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "We introduce LoopArena, a benchmark for evaluating how well one model can guide a separate coding agent through a long-running task.", "canonicalNameSource": "abstract", "decisionReason": "The benchmark has a clear name, defined settings and scoring, and is released with code and data on GitHub, providing a credible public reuse path.", "model": "deepseek-v4-pro", "reviewedAt": "2026-09-05T08:28:38.611381Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates a controller model guiding a fixed coding agent through long-running tasks, scored via contract accuracy, strict success rate, and cost metrics across three complementary settings.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["exact coined title identity tied to benchmark evidence", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We introduce LoopArena, a benchmark for evaluating how well one model can guide a separate coding agent through a long-running task."}, "familyId": "bmf_0b2777f3ca58", "firstSeenAt": "2026-08-31", "heat": null, "id": "bm_looparena_97bc3500", "indexedAt": "2026-08-31T07:22:53.604319Z", "industrySectors": [], "links": {"code": "https://github.com/AMAP-ML/LoopArena", "data": null, "hfPaper": "https://huggingface.co/papers/2608.28281", "paper": "http://arxiv.org/abs/2608.28281v1", "pdf": "https://arxiv.org/pdf/2608.28281v1", "project": null, "report": "http://arxiv.org/abs/2608.28281v1"}, "metrics": [], "motivation": "Loop Engineering is emerging as a practice for organizing development work around coding agents.", "name": "LoopArena", "oneLine": "Evaluates a controller model guiding a fixed coding agent through long-running tasks, scored via contract accuracy, strict success rate, and cost metrics across three complementary settings.", "paperTitle": "LoopArena: Benchmarking Models as Runtime Controllers for Loop Engineering", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "http://arxiv.org/abs/2608.28281v1", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-09-05T06:02:50.019181Z"}, "publications": [], "publishers": [{"name": "DreamX Team", "organizationType": "community", "role": "benchmark-publisher", "sourceUrl": "https://github.com/AMAP-ML/LoopArena"}], "ranking": {"30d": {"confidence": "High", "coverage": 0.85, "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": 30}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": 0.6668488377727539, "value": 87}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.8192394503285512, "value": 101}}, "confidence": "High", "coverage": 0.85, "rank": 5, "score": 69}, "method": "attention-ranking-v13", "rank": 5, "score": 69}, "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": 0.6062554930400308, "value": 87}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.8192394503285512, "value": 101}}, "confidence": "Medium", "coverage": 0.7, "rank": 17, "score": 61}, "method": "attention-ranking-v13", "rank": 17, "score": 61}}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-08-28", "source": {"authors": ["Yi Wang", "Haopeng Zhang", "Chengxiang Huang", "Rui Dai", "Kaikui Liu", "Piotr Koniusz", "Xiangxiang Chu"], "categories": ["cs.AI"], "id": "2608.28281", "title": "LoopArena: Benchmarking Models as Runtime Controllers for Loop Engineering", "type": "arxiv", "url": "http://arxiv.org/abs/2608.28281v1"}, "sourceUpdatedAt": "2026-08-28T12:44:54Z", "topics": ["Agents"], "venueAttempts": [], "whyItMatters": "Separates loop guidance quality from worker capability, addressing a gap in evaluating coding agent orchestration and providing a reproducible protocol for runtime controller comparison."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
