{"record": {"aliases": ["$R^3$-Bench: LLMs Struggle with Resource-Rational Reasoning under Shared Budgets"], "artifacts": [{"id": "arxiv:2608.16033", "kind": "paper", "url": "https://arxiv.org/abs/2608.16033"}, {"id": "gh:nineabyss/r-3-bench", "kind": "repo", "url": "https://github.com/NineAbyss/R-3-Bench"}, {"id": "hf:r-3-bench/r-3-bench", "kind": "dataset", "url": "https://huggingface.co/datasets/R-3-Bench/R-3-Bench"}], "authors": ["Peisong Wang", "Zhiwei Ma", "Bowen Liu", "Feixue Liu", "Aochuan Chen", "Chenyi Zi", "Hongchuan Zeng", "Yuhan Li", "Jia Li"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.16033", "source_url": "https://arxiv.org/abs/2608.16033"}, "categories": ["Code & Software", "Code generation", "General AI", "Reasoning"], "description": {"en": "Evaluates resource-rational reasoning by measuring how models allocate a shared budget across six-problem suites in mathematics, competitive programming, and abstract reasoning, with scoring based on completed problems."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.16033", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.16033", "title": "R3-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["The Hong Kong University of Science and Technology (Guangzhou)", "University of Hong Kong", "Tencent (China)"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.16033", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.16033", "source_url": "https://openalex.org/W7203692504"}, "key": "claire-radar:2608.16033", "modality": null, "name": "R3-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_r-3-bench-llms-struggle-with-resource-rati_fbfbc38c", "origin_source": "arxiv", "recognition_confidence": "0.7", "record_sha256": "b187456f83fbe06d1a79275169f6e82c4009cc27ae5deea3733837ab3bb0e8e2", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-08-25T23:23:46.758372Z", "source_url": "https://arxiv.org/abs/2608.16033"}, "publisher": null, "released": "2026-08-17", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.16033", "source_url": "https://arxiv.org/abs/2608.16033"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-16033", "source": "claire_radar", "source_benchmark_id": "2608.16033", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["$R^3$-Bench: LLMs Struggle with Resource-Rational Reasoning under Shared Budgets"], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Code & Software", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/NineAbyss/R-3-Bench", "githubScope": "benchmark_repo", "githubStars": 3, "hfDailySubmittedAt": null, "hfDataset": "R-3-Bench/R-3-Bench", "hfDatasetDownloads": 311, "hfDatasetLikes": 6, "hfPaperUpvotes": 18, "hfPaperUrl": "https://huggingface.co/papers/2608.16033", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-30", "state": "stale"}, "hfDatasetLikes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-30", "state": "stale"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-30", "state": "stale"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Clear public artifacts and a novel resource-rational framing are likely to draw moderate interest from reasoning and agent evaluation researchers.", "score": 55}, "capabilities": ["Reasoning", "Code generation"], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-25T23:23:46.758372Z", "inputHash": "2d149a6a33a106980178af7ce27e01bd513c59b5924dcff4b418fcc409a5aa12", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "$R^3$-Bench: LLMs Struggle with Resource-Rational Reasoning under Shared Budgets", "canonicalNameSource": "paper_title", "decisionReason": "R3-Bench has a formal name, a stable scoring contract documented in the evaluator, and public code and data on GitHub and Hugging Face enabling reuse and comparison.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-25T23:23:46.758372Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates resource-rational reasoning by measuring how models allocate a shared budget across six-problem suites in mathematics, competitive programming, and abstract reasoning, with scoring based on completed problems.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We introduce $R^3$-Bench, which evaluates six-problem suites under shared budgets across mathematics, competitive programming, and abstract reasoning in tool-free and agentic settings."}, "familyId": "bmf_6d88db5c68c9", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_r-3-bench-llms-struggle-with-resource-rati_fbfbc38c", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/NineAbyss/R-3-Bench", "data": "https://huggingface.co/datasets/R-3-Bench/R-3-Bench", "paper": "https://arxiv.org/abs/2608.16033", "pdf": "https://arxiv.org/pdf/2608.16033", "project": null, "report": "https://arxiv.org/abs/2608.16033"}, "metrics": [], "motivation": "In cognitive science, resource rationality asks how an agent should allocate limited computation to maximize expected value.", "name": "R3-Bench", "oneLine": "Evaluates resource-rational reasoning by measuring how models allocate a shared budget across six-problem suites in mathematics, competitive programming, and abstract reasoning, with scoring based on completed problems.", "paperTitle": "$R^3$-Bench: LLMs Struggle with Resource-Rational Reasoning under Shared Budgets", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.16033", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "ranking": {"90d": {"confidence": "High", "coverage": 1.0, "datasetDownloadRank": 35, "datasetRankPopulation": 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": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": 0.18504537701418716, "value": 3}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": 0.5914140254000603, "value": 311}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.5202664955757326, "value": 18}}, "confidence": "High", "coverage": 1.0, "rank": 176, "score": 36}, "method": "attention-ranking-v13", "rank": 176, "score": 36}}, "readiness": "Runnable", "recognitionConfidence": 0.7, "relation": "introduces", "releasedAt": "2026-08-17", "source": {"authors": ["Peisong Wang", "Zhiwei Ma", "Bowen Liu", "Feixue Liu", "Aochuan Chen", "Chenyi Zi", "Hongchuan Zeng", "Yuhan Li", "Jia Li"], "categories": ["cs.CL"], "id": "2608.16033", "title": "$R^3$-Bench: LLMs Struggle with Resource-Rational Reasoning under Shared Budgets", "type": "arxiv", "url": "https://arxiv.org/abs/2608.16033"}, "sourceUpdatedAt": "2026-08-17T02:59:58Z", "topics": ["Reasoning"], "venueAttempts": [], "whyItMatters": "Addresses the gap between demonstrated single-problem competence and shared-budget allocation, providing a fixed dataset and protocol for studying resource allocation in LLMs."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
