{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.23564", "kind": "paper", "url": "https://arxiv.org/abs/2608.23564v1"}], "authors": ["Deyao Hong", "Yizhe Chi", "Wenyi Li", "Xiaoqiu Wang", "Mingju Gao", "Kaisen Yang", "Bingxiang He", "Youjie Zheng", "Calvin Xiao", "Qinhuai Na"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.23564", "source_url": "https://arxiv.org/abs/2608.23564"}, "categories": ["Language & Knowledge", "Software & AI Compute", "cs.CL"], "description": {"en": "SWE Refactor Bench is a benchmark for evaluating coding agents on whole-repository stack migrations. It comprises 20 migrations covering 4 types of technical debt, with a three-stage evaluation protocol measuring migration completeness and behavioral correctness: Migration Audit, Behavioral Tests, and Agentic Verification."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.23564v1", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.23564v1", "title": "SWE Refactor Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Tsinghua University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.23564", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.23564", "source_url": "https://openalex.org/W7204204864"}, "key": "claire-radar:2608.23564", "modality": null, "name": "SWE Refactor 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_swe-refactor-bench_45bb76b0", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "64b759e828573ba5d3e5fb60f2385f3c7c5c35e0030872bb79777420adf554f6", "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/2608.23564v1"}, "publisher": null, "released": "2026-08-24", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.23564", "source_url": "https://arxiv.org/abs/2608.23564v1"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-23564", "source": "claire_radar", "source_benchmark_id": "2608.23564", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Software & AI Compute"], "area": "Language & Knowledge", "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/2608.23564", "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 the trending area of AI coding agents with a robust evaluation, likely attracting significant interest from developers and researchers.", "score": 60}, "capabilities": [], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-27T04:22:25.189955Z", "inputHash": "724748fbcbf6cfeb487c90ef3343c92676cf83df84f0720b6b9750afca9fe00f", "model": "deepseek-v4-flash"}, "curation": {"canonicalNameEvidence": "We introduce SWE Refactor Bench, a benchmark comprising 20 whole-repository migrations", "canonicalNameSource": "abstract", "decisionReason": "Formally named benchmark with a rigorous multi-stage protocol and results from frontier models.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-27T04:18:56.318436Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SWE Refactor Bench is a benchmark for evaluating coding agents on whole-repository stack migrations. It comprises 20 migrations covering 4 types of technical debt, with a three-stage evaluation protocol measuring migration completeness and behavioral correctness: Migration Audit, Behavioral Tests, and Agentic Verification.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "To address this problem, we introduce SWE Refactor Bench, a benchmark comprising 20 whole-repository migrations, covering 4 kinds of technical debt."}, "familyId": "bmf_e82c050bd0ea", "firstSeenAt": "2026-08-25", "heat": null, "id": "bm_swe-refactor-bench_45bb76b0", "indexedAt": "2026-08-25T16:21:05.134052Z", "industrySectors": ["Software & Cloud"], "links": {"code": null, "data": null, "paper": "http://arxiv.org/abs/2608.23564v1", "pdf": "https://arxiv.org/pdf/2608.23564v1", "project": null, "report": "http://arxiv.org/abs/2608.23564v1"}, "metrics": [], "motivation": "Modern software systems accumulate technical debt over decades of development, which makes migration expensive and largely manual.", "name": "SWE Refactor Bench", "oneLine": "SWE Refactor Bench is a benchmark for evaluating coding agents on whole-repository stack migrations. It comprises 20 migrations covering 4 types of technical debt, with a three-stage evaluation protocol measuring migration completeness and behavioral correctness: Migration Audit, Behavioral Tests, and Agentic Verification.", "paperTitle": "SWE Refactor Bench: Can Coding Agents Complete a Long-Horizon, Whole-Repository Stack Migration?", "primaryDomain": "Software & AI Compute", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "http://arxiv.org/abs/2608.23564v1", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-25T16:21:05.134052Z"}, "publications": [], "publishers": [{"name": "SWE Refactor Bench Team", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "http://arxiv.org/abs/2608.23564v1"}], "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": 1.0, "relation": "introduces", "releasedAt": "2026-08-24", "source": {"authors": ["Deyao Hong", "Yizhe Chi", "Wenyi Li", "Xiaoqiu Wang", "Mingju Gao", "Kaisen Yang", "Bingxiang He", "Youjie Zheng", "Calvin Xiao", "Qinhuai Na"], "categories": ["cs.CL", "cs.AI", "cs.SE"], "id": "2608.23564", "title": "SWE Refactor Bench: Can Coding Agents Complete a Long-Horizon, Whole-Repository Stack Migration?", "type": "arxiv", "url": "http://arxiv.org/abs/2608.23564v1"}, "sourceUpdatedAt": "2026-08-24T17:59:04Z", "topics": ["cs.CL"], "venueAttempts": [], "whyItMatters": "The benchmark addresses the gap in evaluating code migration beyond mere test passing, preventing shortcut solutions. It provides a rigorous testbed for developing reliable coding agents for long-horizon refactoring tasks, with findings on agent capability across migration categories."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": [{"key": "model-reports:swe_refactor_bench", "name": "SWE Refactor Bench", "relation": "equivalent", "slug": "swe_refactor_bench", "source": "model_reports"}]}
