Elamur Research⌕
☰
home / papers / 2025 / CodeTree: Agent-guided Tree Search for Code Generation with Large Language Models

CodeTree: Agent-guided Tree Search for Code Generation with Large Language Models

Paper · 2025 · Jierui Li, Hung Le, Yingbo Zhou et al.

Summary

Hard programming tasks have a huge space of possible solutions. Older methods either sample many independent candidates or repeatedly refine one early attempt, which can trap them in a local optimum. CodeTree addresses this by organizing code generation as a tree search run by cooperating LLM agents. A Thinker agent proposes several natural-language strategies, a Solver turns each into code, and a Debugger revises failing programs. A Critic agent uses both test execution results and its own LLM judgments to score nodes and decide whether to expand, abort or accept a branch. This lets the search avoid redundant or dead-end paths.

The authors tested the framework on 7 code generation benchmarks, from beginner exercises to competition problems, and adapted it to repository-level patching on SWEBench. With GPT-4o, it beat strong baselines and matched or exceeded o1-preview while often using fewer tokens. The work suggests that structured, agent-directed exploration is a more efficient use of a fixed generation budget than brute-force sampling or single-path refinement.

Key points

  • Using GPT-4o as its base model and a generation budget of 20 samples, CodeTree achieves pass@1 scores of 95.1% (HumanEval), 98.7% (MBPP) and 43.0% (CodeContests).
  • With a budget of 30, CodeTree scored 49.1% on CodeContests using about 4.8k tokens per problem. o1-preview scored 46.6% with about 7.1k tokens. CodeTree also led on MBPP+ (82.8% vs 80.2%) and HumanEval+ (89.6% vs 89.0%).
  • On SWEBench with GPT-4o-mini, the authors built on the pipeline of Xia et al. (2024) and recast strategy generation as context retrieval. CodeTree resolved 27.6% of issues, compared with 24.2% for the base pipeline, 23.6% with CoT and 25.3% with Reflexion.
  • The framework needs strong instruction-following models; 8B-scale models produced unreliable Critic feedback. The method also optimizes only functional correctness, not readability or efficiency.
  • Code is released at github.com/SalesforceAIResearch/CodeTree.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓The University of Texas at AustinAustin · United States109 works · 56,942 citations🏢SalesforceSan Francisco · United States56 works · 17,867 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksInformation Systems · 1,437 worksComputer Science Applications · 102 works

🧭 Topics 1Research subjects and languages the work is about.Code & programming languages · 129 works

Cite

Cite as Jierui Li, Hung Le, Yingbo Zhou, Caiming Xiong, Silvio Savarese and Doyen Sahoo. 2025. CodeTree: Agent-guided Tree Search for Code Generation with Large Language Models. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2025.naacl-long.189

@inproceedings{li2025codetree,
  title = {{CodeTree: Agent-guided Tree Search for Code Generation with Large Language Models}},
  author = {Jierui Li, Hung Le, Yingbo Zhou, Caiming Xiong, Silvio Savarese and Doyen Sahoo},
  booktitle = {Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)},
  year = {2025},
  doi = {10.18653/v1/2025.naacl-long.189},
  url = {https://doi.org/10.18653/v1/2025.naacl-long.189}
}

References

46 references, 22 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774.
  2. Jacob Austin, Augustus Odena, Maxwell Nye, Maarten Bosma, Henryk Michalewski, David Dohan, Ellen Jiang, Carrie Cai, Michael Terry, Quoc Le, et al. 2021. Program synthesis with large language models. arXiv preprint arXiv:2108.07732.
  3. Angelica Chen, Jérémy Scheurer, Tomasz Korbak, Jon Ander Campos, Jun Shern Chan, Samuel R Bowman, Kyunghyun Cho, and Ethan Perez. 2023a. Improving code generation by training with natural language feedback. arXiv preprint arXiv:2303.16749. find
  4. Bei Chen, Fengji Zhang, Anh Nguyen, Daoguang Zan, Zeqi Lin, Jian-Guang Lou, and Weizhu Chen. 2023b. Codet: Code generation with generated tests. In The Eleventh International Conference on Learning Representations. find
  5. Jingchang Chen, Hongxuan Tang, Zheng Chu, Qianglong Chen, Zekun Wang, Ming Liu, and Bing Qin. 2024. Divide-and-conquer meets consensus: Unleashing the power of functions in code generation. arXiv preprint arXiv:2405.20092. — [arXiv:2405.20092](https://arxiv.org/abs/2405.20092) find
  6. Mark Chen, Jerry Tworek, Heewoo Jun, Qiming Yuan, Henrique Ponde de Oliveira Pinto, Jared Kaplan, Harri Edwards, Yuri Burda, Nicholas Joseph, Greg Brockman, et al. 2021. Evaluating large language models trained on code. arXiv preprint arXiv:2107.03374.
  7. Weize Chen, Yusheng Su, Jingwei Zuo, Cheng Yang, Chenfei Yuan, Chi-Min Chan, Heyang Yu, Yaxi Lu, Yi-Hsin Hung, Chen Qian, et al. 2023c. Agentverse: Facilitating multi-agent collaboration and exploring emergent behaviors. In The Twelfth International Conference on Learning Representations.
  8. Xinyun Chen, Maxwell Lin, Nathanael Schärli, and Denny Zhou. 2023d. Teaching large language models to self-debug. arXiv preprint arXiv:2304.05128. find
  9. Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, Parker Schuh, Kensen Shi, Sasha Tsvyashchenko, Joshua Maynez, Abhishek Rao, Parker Barnes, Yi Tay, Noam Shazeer, Vinodkumar Prabhakaran, Emi… — PaLM: Scaling Language Modeling with Pathways
  10. Yihong Dong, Xue Jiang, Zhi Jin, and Ge Li. 2023. Self-collaboration code generation via chatgpt. arXiv preprint arXiv:2304.07590. find
  11. Suriya Gunasekar, Yi Zhang, Jyoti Aneja, Caio César Teodoro Mendes, Allie Del Giorno, Sivakanth Gopi, Mojan Javaheripi, Piero Kauffmann, Gustavo de Rosa, Olli Saarikivi, et al. 2023. Textbooks are all you need. arXiv preprint arXiv:2306.11644. find
  12. Dan Hendrycks, Steven Basart, Saurav Kadavath, Mantas Mazeika, Akul Arora, Ethan Guo, Collin Burns, Samir Puranik, Horace He, Dawn Song, and Jacob Steinhardt. 2021. Measuring coding challenge competence with apps. NeurIPS.
  13. Sirui Hong, Xiawu Zheng, Jonathan Chen, Yuheng Cheng, Jinlin Wang, Ceyao Zhang, Zili Wang, Steven Ka Shing Yau, Zijuan Lin, Liyang Zhou, et al. 2023. Metagpt: Meta programming for multi-agent collaborative framework. arXiv preprint arXiv:2308.00352.
  14. Dong Huang, Qingwen Bu, Jie M Zhang, Michael Luck, and Heming Cui. 2023. Agentcoder: Multi-agent-based code generation with iterative testing and optimisation. arXiv preprint arXiv:2312.13010. find
  15. Carlos E Jimenez, John Yang, Alexander Wettig, Shunyu Yao, Kexin Pei, Ofir Press, and Karthik R Narasimhan. 2024. SWE-bench: Can language models resolve real-world github issues? In The Twelfth International Conference on Learning Representations.
  16. Jing Yu Koh, Stephen McAleer, Daniel Fried, and Ruslan Salakhutdinov. 2024. Tree search for language model agents. arXiv preprint arXiv:2407.01476. find
  17. Hung Le, Hailin Chen, Amrita Saha, Akash Gokul, Doyen Sahoo, and Shafiq Joty. 2024. Codechain: Towards modular code generation through chain of self-revisions with representative sub-modules. In The Twelfth International Conference on Learning Representations. find
  18. Hung Le, Yue Wang, Akhilesh Deepak Gotmare, Silvio Savarese, and Steven Chu Hong Hoi. 2022. Coderl: Mastering code generation through pretrained models and deep reinforcement learning. Advances in Neural Information Processing Systems, 35:21314–21328.
  19. Guohao Li, Hasan Hammoud, Hani Itani, Dmitrii Khizbullin, and Bernard Ghanem. 2023a. Camel: Communicative agents for" mind" exploration of large language model society. Advances in Neural Information Processing Systems, 36:51991–52008.
  20. Raymond Li, Loubna Ben Allal, Yangtian Zi, Niklas Muennighoff, Denis Kocetkov, Chenghao Mou, Marc Marone, Christopher Akiki, Jia Li, Jenny Chim, et al. 2023b. Starcoder: may the source be with you! arXiv preprint arXiv:2305.06161. find
  21. Yujia Li, David Choi, Junyoung Chung, Nate Kushman, Julian Schrittwieser, Rémi Leblond, Tom Eccles, James Keeling, Felix Gimeno, Agustin Dal Lago, et al. 2022. Competition-level code generation with alphacode. arXiv preprint arXiv:2203.07814.
  22. Jiawei Liu, Chunqiu Steven Xia, Yuyao Wang, and Lingming Zhang. 2023. Is your code generated by chat-GPT really correct? rigorous evaluation of large language models for code generation. In Thirty-seventh Conference on Neural Information Processing Systems. find
  23. Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegreffe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, et al. 2023. Self-refine: Iterative refinement with self-feedback. arXiv preprint arXiv:2303.17651. find
  24. Jeevana Priya Inala, Chenglong Wang, Mei Yang, Andres Codas, Mark Encarnación, Shuvendu K Lahiri, Madanlal Musuvathi, and Jianfeng Gao. 2022. Fault-aware neural code rankers. In Advances in Neural Information Processing Systems. find
  25. Ansong Ni, Srini Iyer, Dragomir Radev, Veselin Stoyanov, Wen-tau Yih, Sida Wang, and Xi Victoria Lin. 2023. Lever: Learning to verify language-to-code generation with execution. In International Conference on Machine Learning, pages 26106–26128. PMLR.
  26. Md Ashraful Islam, Mohammed Eunus Ali, and Md Rizwan Parvez. 2024. Mapcoder: Multi-agent code generation for competitive problem solving. arXiv preprint arXiv:2405.11403. find
  27. Erik Nijkamp, Hiroaki Hayashi, Caiming Xiong, Silvio Savarese, and Yingbo Zhou. 2023. Codegen2: Lessons for training llms on programming and natural languages. arXiv preprint arXiv:2305.02309. find
  28. Theo X Olausson, Jeevana Priya Inala, Chenglong Wang, Jianfeng Gao, and Armando Solar-Lezama. 2023. Demystifying gpt self-repair for code generation. arXiv preprint arXiv:2306.09896. find
  29. Long Ouyang, Jeff Wu, Xu Jiang, Diogo Almeida, Carroll L Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. 2022. Training language models to follow instructions with human feedback. arXiv preprint arXiv:2203.02155.
  30. Chen Qian, Wei Liu, Hongzhang Liu, Nuo Chen, Yufan Dang, Jiahao Li, Cheng Yang, Weize Chen, Yusheng Su, Xin Cong, et al. 2024. Chatdev: Communicative agents for software development. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), p…
  31. Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, et al. 2023. Code llama: Open foundation models for code. arXiv preprint arXiv:2308.12950.
  32. Freda Shi, Daniel Fried, Marjan Ghazvininejad, Luke Zettlemoyer, and Sida I. Wang. 2022. Natural language to code translation with execution. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 3533–3546, Abu Dhabi, United Arab Emirates. Association for C… find
  33. Noah Shinn, Federico Cassano, Beck Labash, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. 2023. Reflexion: Language agents with verbal reinforcement learning. Preprint, arXiv:2303.11366.
  34. Jialin Song, Jonathan Raiman, and Bryan Catanzaro. 2024. Effective large language model debugging with best-first tree search. arXiv preprint arXiv:2407.19055. — [arXiv:2407.19055](https://arxiv.org/abs/2407.19055) find
  35. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023a. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971.
  36. Codet5+: Open code large language models for code understanding and generation. arXiv preprint arXiv:2305.07922. find
  37. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems, 35:24824–24837.
  38. Sean Welleck, Ximing Lu, Peter West, Faeze Brahman, Tianxiao Shen, Daniel Khashabi, and Yejin Choi. 2023. Generating sequences by learning to self-correct. In The Eleventh International Conference on Learning Representations.
  39. Chunqiu Steven Xia, Yinlin Deng, Soren Dunn, and Lingming Zhang. 2024. Agentless: Demystifying llm-based software engineering agents. arXiv preprint arXiv:2407.01489. find
  40. Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Tom Griffiths, Yuan Cao, and Karthik Narasimhan. 2024. Tree of thoughts: Deliberate problem solving with large language models. Advances in Neural Information Processing Systems, 36. find
  41. Kechi Zhang, Zhuo Li, Jia Li, Ge Li, and Zhi Jin. 2023a. Self-edit: Fault-aware code editor for code generation. arXiv preprint arXiv:2305.04087. find
  42. Tianyi Zhang, Tao Yu, Tatsunori Hashimoto, Mike Lewis, Wen-tau Yih, Daniel Fried, and Sida Wang. 2023b. Coder reviewer reranking for code generation. In International Conference on Machine Learning, pages 41832–41846. PMLR. find
  43. Denny Zhou, Nathanael Schärli, Le Hou, Jason Wei, Nathan Scales, Xuezhi Wang, Dale Schuurmans, Claire Cui, Olivier Bousquet, Quoc V Le, and Ed H. Chi. 2023. Least-to-most prompting enables complex reasoning in large language models. In The Eleventh International Conference on Learning Representation…
  44. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. 2023b. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288.
  45. Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc Le, Ed Chi, Sharan Narang, Aakanksha Chowdhery, and Denny Zhou. 2022. Self-consistency improves chain of thought reasoning in language models. arXiv preprint arXiv:2203.11171.
  46. Yue Wang, Hung Le, Akhilesh Deepak Gotmare, Nghi DQ Bui, Junnan Li, and Steven CH Hoi. 2023. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/li-2025-codetree-agent-guided-tree-search/