Year: 2025
Type: conference-paper
Abstract: Large language models (LLMs) have advanced rapidly from conversational problem solving to addressing real-world tasks involving tool use, such as software engineering (SWE).Recent LLM-powered SWE systems, such as Ope-nAI Codex and Cursor, have offered end-to-end automation of the software development process.However, building effective SWE agents remains challenging due to the lack of highquality training data and reliable test-time evaluation.To address this issue, we present SWE-DEV, an SWE agent built upon open-source LLMs, with a focus on training and inference scaling.For training scaling, we develop a robust pipeline to synthesize test cases and scale up agent trajectories to construct the training data.For inference scaling, we increase the interaction budget within a single run to enable further thinking within one independent attempt.Experiments on the SWE-bench-Verified benchmark show that the SWE-DEV models can achieve top performance among all open SWE agents.Specifically, the resolve rate of our 7B and 32B models reach 23.4% and 36.6%,respectively, outperforming state-of-the-art open-source models.All code, models, and datasets are publicly available at https://github.com/THUDM/SWE-Dev. OpenAI. 2024b.Learning to reason with llms.https://openai.com/index/ learning-to-reason-with-llms.
Authors: Haoran Wang, Zhenyu Hou, Yao Wei, Jie Tang, Yuxiao Dong
Institution: Tsinghua University
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4412888747