Satori-SWE: Evolutionary Test-Time Scaling for Sample-Efficient Software Engineering

TL;DR

EvoScale is a sample-efficient evolutionary method that improves smaller language models' performance on software engineering tasks by self-evolving outputs through reinforcement learning, reducing the need for extensive sampling.

Contribution

It introduces EvoScale, a novel reinforcement learning-based approach that enables smaller language models to self-improve and match larger models' performance in software engineering tasks.

Findings

32B model matches or exceeds 100B models on SWE-Bench.

EvoScale reduces sampling requirements significantly.

Self-evolving models achieve high accuracy with fewer resources.

Abstract

Language models (LMs) perform well on standardized coding benchmarks but struggle with real-world software engineering tasks such as resolving GitHub issues in SWE-Bench, especially when model parameters are less than 100B. While smaller models are preferable in practice due to their lower computational cost, improving their performance remains challenging. Existing approaches primarily rely on supervised fine-tuning (SFT) with high-quality data, which is expensive to curate at scale. An alternative is test-time scaling: generating multiple outputs, scoring them using a verifier, and selecting the best one. Although effective, this strategy often requires excessive sampling and costly scoring, limiting its practical application. We propose Evolutionary Test-Time Scaling (EvoScale), a sample-efficient method that treats generation as an evolutionary process. By iteratively refining outputs via selection and mutation, EvoScale shifts the output distribution toward higher-scoring regions, reducing the number of samples needed to find correct solutions. To reduce the overhead from repeatedly sampling and selection, we train the model to self-evolve using reinforcement learning (RL). Rather than relying on external verifiers at inference time, the model learns to self-improve the scores of its own generations across iterations. Evaluated on SWE-Bench-Verified, EvoScale enables our 32B model, Satori-SWE-32B, to match or exceed the performance of models with over 100B parameters while using a few samples. Code, data, and models will be fully open-sourced.