{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2609.04611", "kind": "paper", "url": "https://arxiv.org/abs/2609.04611"}], "authors": ["Quan Shi", "Keshav Dhandhania", "Karthik Narasimhan", "Victor Barres"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2609.04611", "source_url": "https://arxiv.org/abs/2609.04611"}, "categories": ["Agents", "Agents & Tool Use", "General AI"], "description": {"en": "A benchmark where a developer agent builds a complete customer-service agent from business records, client requirements, a production API, inherited codebase, and cost/model limits, scored by deployment against held-out simulated users across 53 tasks in four domains."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2609.04611", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2609.04611", "title": "$\\tau^\\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Princeton University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2609.04611", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2609.04611", "source_url": "https://openalex.org/W7211915101"}, "key": "claire-radar:2609.04611", "modality": null, "name": "$\\tau^\\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "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_tau-tau-bench-an-environment-for-end-to-en_3d823aef", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "f1480bc782aef9a53fd9cc07d7d3796c547cd9aafc3922098f02440bb764a55f", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-09-10T09:02:48.332472Z", "source_url": "https://arxiv.org/abs/2609.04611"}, "publisher": null, "released": "2026-09-04", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2609.04611", "source_url": "https://arxiv.org/abs/2609.04611"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2609-04611", "source": "claire_radar", "source_benchmark_id": "2609.04611", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Agents & Tool Use", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/sierra-research/hyper-tau-bench", "githubScope": "benchmark_repo", "githubStars": 34, "hfDailySubmittedAt": "2026-09-07T00:00:00.000Z", "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 10, "hfPaperUrl": "https://huggingface.co/papers/2609.04611", "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": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "The novel framing of agent construction as a task and the stark 23.9% pass rate for the strongest configuration are likely to draw interest, though the specialized focus on customer-service agent development may narrow immediate appeal.", "score": 35}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-10T09:02:48.332472Z", "inputHash": "933219932c7bdd8c8d60c576b3df369794f1514896c840243a0ca743083b059a", "model": "deepseek-v4-pro"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "$\\tau^\\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction", "canonicalNameSource": "paper_title", "decisionReason": "The benchmark defines a repeatable evaluation object (agent construction) with a clear scoring contract (deployment against held-out simulated users) and provides a public path through the paper and stated aim for measurable reuse.", "evaluationEvidence": {"reuse": "We aim for $\\tau^\\tau$-bench to turn the work of cooperative agent building into a measurable target for coding agents.", "scoring": "From these it must deliver a complete customer-service agent, scored by deploying that agent against held-out simulated users.", "task": "A developer agent is given the records a business actually keeps, a client who holds requirements, a production API that operations must run through, a codebase to inherit, and limits on serving cost and models."}, "model": "deepseek-v4-pro", "reviewedAt": "2026-09-10T09:02:48.332472Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "A benchmark where a developer agent builds a complete customer-service agent from business records, client requirements, a production API, inherited codebase, and cost/model limits, scored by deployment against held-out simulated users across 53 tasks in four domains.", "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"], "snippet": "We introduce $\\tau^\\tau$-bench (pronounced hyper-tau-bench), a benchmark that makes agent construction the task."}, "familyId": "bmf_83c752c26b32", "firstSeenAt": "2026-09-07", "heat": null, "id": "bm_tau-tau-bench-an-environment-for-end-to-en_3d823aef", "indexedAt": "2026-09-07T06:22:02.605340Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2609.04611", "pdf": "https://arxiv.org/pdf/2609.04611", "project": null, "report": "https://arxiv.org/abs/2609.04611"}, "metrics": [], "motivation": "LLM agents are rapidly becoming production software, deployed to handle customer service, adjudicate disputes, and operate internal systems.", "name": "$\\tau^\\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction", "oneLine": "A benchmark where a developer agent builds a complete customer-service agent from business records, client requirements, a production API, inherited codebase, and cost/model limits, scored by deployment against held-out simulated users across 53 tasks in four domains.", "paperTitle": "$\\tau^\\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2609.04611", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-09-07T06:22:02.605340Z"}, "publications": [], "ranking": {"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": 0.47457505484152435, "value": 34}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.42369516881263364, "value": 10}}, "confidence": "Medium", "coverage": 0.7, "rank": 87, "score": 47}, "method": "attention-ranking-v13", "rank": 87, "score": 47}}, "readiness": "Paper only", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-09-04", "source": {"authors": ["Quan Shi", "Keshav Dhandhania", "Karthik Narasimhan", "Victor Barres"], "categories": ["cs.AI"], "id": "2609.04611", "title": "$\\tau^\\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction", "type": "arxiv", "url": "https://arxiv.org/abs/2609.04611"}, "sourceUpdatedAt": "2026-09-04T01:23:47Z", "topics": ["Agents"], "venueAttempts": [], "whyItMatters": "It turns the work of cooperative agent building into a measurable target for coding agents, evaluating whether AI systems can deliver production agents under real client-engagement conditions."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
