{"paper":{"title":"Language Agent Tree Search Unifies Reasoning Acting and Planning in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Language models use Monte Carlo tree search with self-reflections to plan and act as agents.","cross_cats":["cs.CL","cs.CV","cs.LG"],"primary_cat":"cs.AI","authors_text":"Andy Zhou, Haohan Wang, Kai Yan, Michal Shlapentokh-Rothman, Yu-Xiong Wang","submitted_at":"2023-10-06T17:55:11Z","abstract_excerpt":"While language models (LMs) have shown potential across a range of decision-making tasks, their reliance on simple acting processes limits their broad deployment as autonomous agents. In this paper, we introduce Language Agent Tree Search (LATS) -- the first general framework that synergizes the capabilities of LMs in reasoning, acting, and planning. By leveraging the in-context learning ability of LMs, we integrate Monte Carlo Tree Search into LATS to enable LMs as agents, along with LM-powered value functions and self-reflections for proficient exploration and enhanced decision-making. A key"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"LATS achieves state-of-the-art pass@1 accuracy (92.7%) for programming on HumanEval with GPT-4 and demonstrates gradient-free performance (average score of 75.9) comparable to gradient-based fine-tuning for web navigation on WebShop with GPT-3.5.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That language models can reliably serve as value functions and self-reflectors within Monte Carlo Tree Search using only in-context learning and external environment feedback, without needing task-specific training.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"LATS integrates Monte Carlo Tree Search with language models using in-context learning, value functions, and self-reflection to achieve 92.7% pass@1 on HumanEval and competitive web navigation performance.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Language models use Monte Carlo tree search with self-reflections to plan and act as agents.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"5892c298ba859d75ea61aa2dc47d8bb6d764df8491c2e8ee9aa0a1a23f3a91a9"},"source":{"id":"2310.04406","kind":"arxiv","version":3},"verdict":{"id":"58a4c1bb-024c-4319-b169-45e58f2a7e48","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T23:21:58.759085Z","strongest_claim":"LATS achieves state-of-the-art pass@1 accuracy (92.7%) for programming on HumanEval with GPT-4 and demonstrates gradient-free performance (average score of 75.9) comparable to gradient-based fine-tuning for web navigation on WebShop with GPT-3.5.","one_line_summary":"LATS integrates Monte Carlo Tree Search with language models using in-context learning, value functions, and self-reflection to achieve 92.7% pass@1 on HumanEval and competitive web navigation performance.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That language models can reliably serve as value functions and self-reflectors within Monte Carlo Tree Search using only in-context learning and external environment feedback, without needing task-specific training.","pith_extraction_headline":"Language models use Monte Carlo tree search with self-reflections to plan and act as agents."},"references":{"count":13,"sample":[{"doi":"","year":null,"title":"Graph of thoughts: Solving elaborate problems with large language models","work_id":"34ed66c5-13ac-4799-b958-8e57d3de4705","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","ref_index":2,"cited_arxiv_id":"2107.03374","is_internal_anchor":true},{"doi":"","year":null,"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","ref_index":3,"cited_arxiv_id":"2110.14168","is_internal_anchor":true},{"doi":"","year":null,"title":"Mastering Diverse Domains through World Models","work_id":"6aeb260f-8c7c-4f9c-b98b-067cd7c59acd","ref_index":4,"cited_arxiv_id":"2301.04104","is_internal_anchor":true},{"doi":"","year":null,"title":"Reason for fu- ture, act for now: A principled framework for au- tonomous LLM agents with provable sample efficiency","work_id":"ba51c107-b4c4-45ab-a1ae-5164af8a0587","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":13,"snapshot_sha256":"b9d8f897f7b090517e4fe11ad6f64a7325bfeeedb790d403733f144fd63baf18","internal_anchors":7},"formal_canon":{"evidence_count":2,"snapshot_sha256":"9e8bc7682508675408b709a73a7f8b476de14435e933a514458a54ed2e236fc0"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}