SWE-Master: Unleashing the Potential of Software Engineering Agents via Post-Training

TL;DR

SWE-Master is a comprehensive post-training framework that significantly enhances the capabilities of software engineering agents through systematic optimization, achieving state-of-the-art results on standard benchmarks.

Contribution

It introduces SWE-Master, a fully reproducible framework for developing effective software engineering agents using systematic optimization and post-training techniques.

Findings

Achieves a 61.4% resolve rate on SWE-bench Verified with Qwen2.5-Coder-32B.

Reaches 70.8% at TTS@8 with test-time scaling and environment feedback.

Outperforms existing open-source baselines in software engineering tasks.

Abstract

In this technical report, we present SWE-Master, an open-source and fully reproducible post-training framework for building effective software engineering agents. SWE-Master systematically explores the complete agent development pipeline, including teacher-trajectory synthesis and data curation, long-horizon SFT, RL with real execution feedback, and inference framework design. Starting from an open-source base model with limited initial SWE capability, SWE-Master demonstrates how systematical optimization method can elicit strong long-horizon SWE task solving abilities. We evaluate SWE-Master on SWE-bench Verified, a standard benchmark for realistic software engineering tasks. Under identical experimental settings, our approach achieves a resolve rate of 61.4\% with Qwen2.5-Coder-32B, substantially outperforming existing open-source baselines. By further incorporating test-time scaling~(TTS) with LLM-based environment feedback, SWE-Master reaches 70.8\% at TTS@8, demonstrating a strong performance potential. SWE-Master provides a practical and transparent foundation for advancing reproducible research on software engineering agents. The code is available at https://github.com/RUCAIBox/SWE-Master.