Trae Agent: An LLM-based Agent for Software Engineering with Test-time Scaling

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Trae Research Team, Gao, Pengfei, Tian, Zhao, Meng, Xiangxin, Wang, Xinchen, Hu, Ruida, Xiao, Yuanan, Liu, Yizhou, Zhang, Zhao, Chen, Junjie, Gao, Cuiyun, Lin, Yun, Xiong, Yingfei, Peng, Chao, Liu, Xia
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915419567161344
author Trae Research Team
Gao, Pengfei
Tian, Zhao
Meng, Xiangxin
Wang, Xinchen
Hu, Ruida
Xiao, Yuanan
Liu, Yizhou
Zhang, Zhao
Chen, Junjie
Gao, Cuiyun
Lin, Yun
Xiong, Yingfei
Peng, Chao
Liu, Xia
author_facet Trae Research Team
Gao, Pengfei
Tian, Zhao
Meng, Xiangxin
Wang, Xinchen
Hu, Ruida
Xiao, Yuanan
Liu, Yizhou
Zhang, Zhao
Chen, Junjie
Gao, Cuiyun
Lin, Yun
Xiong, Yingfei
Peng, Chao
Liu, Xia
contents Software issue resolution is a critical challenge in software engineering and has garnered increasing attention in recent years. With the rapid advancement of large language models (LLMs), substantial progress has been made in addressing real-world software engineering tasks. Recent studies have introduced ensemble reasoning techniques to enhance the performance of LLM-based issue resolution. However, existing prompting-based methods still face limitations in effectively exploring large ensemble spaces and lack the capacity for repository-level understanding, both of which constrain their overall effectiveness. In this paper, we propose Trae Agent, the first agent-based ensemble reasoning approach for repository-level issue resolution. Trae Agent formulates our goal as an optimal solution search problem and addresses two key challenges, i.e., large ensemble spaces and repository-level understanding, through modular agents for generation, pruning, and selection. We conduct extensive experiments using three leading LLMs on the widely-adopted SWE-bench benchmark, comparing Trae Agent against four state-of-the-art ensemble reasoning techniques. Experimental results demonstrate that Trae Agent consistently achieves superior performance, with an average improvement of 10.22% over all baselines in terms of Pass@1. Trae Agent has achieved first place on the SWE-bench Verified leaderboard, with a notable Pass@1 score of 75.20%. We are pleased to release Trae Agent as an open-source project to support the research community, with all resources available at https://github.com/bytedance/trae-agent.
format Preprint
id arxiv_https___arxiv_org_abs_2507_23370
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Trae Agent: An LLM-based Agent for Software Engineering with Test-time Scaling
Trae Research Team
Gao, Pengfei
Tian, Zhao
Meng, Xiangxin
Wang, Xinchen
Hu, Ruida
Xiao, Yuanan
Liu, Yizhou
Zhang, Zhao
Chen, Junjie
Gao, Cuiyun
Lin, Yun
Xiong, Yingfei
Peng, Chao
Liu, Xia
Software Engineering
Artificial Intelligence
Software issue resolution is a critical challenge in software engineering and has garnered increasing attention in recent years. With the rapid advancement of large language models (LLMs), substantial progress has been made in addressing real-world software engineering tasks. Recent studies have introduced ensemble reasoning techniques to enhance the performance of LLM-based issue resolution. However, existing prompting-based methods still face limitations in effectively exploring large ensemble spaces and lack the capacity for repository-level understanding, both of which constrain their overall effectiveness. In this paper, we propose Trae Agent, the first agent-based ensemble reasoning approach for repository-level issue resolution. Trae Agent formulates our goal as an optimal solution search problem and addresses two key challenges, i.e., large ensemble spaces and repository-level understanding, through modular agents for generation, pruning, and selection. We conduct extensive experiments using three leading LLMs on the widely-adopted SWE-bench benchmark, comparing Trae Agent against four state-of-the-art ensemble reasoning techniques. Experimental results demonstrate that Trae Agent consistently achieves superior performance, with an average improvement of 10.22% over all baselines in terms of Pass@1. Trae Agent has achieved first place on the SWE-bench Verified leaderboard, with a notable Pass@1 score of 75.20%. We are pleased to release Trae Agent as an open-source project to support the research community, with all resources available at https://github.com/bytedance/trae-agent.
title Trae Agent: An LLM-based Agent for Software Engineering with Test-time Scaling
topic Software Engineering
Artificial Intelligence
url https://arxiv.org/abs/2507.23370