SWE Refactor Bench: Can Coding Agents Complete a Long-Horizon, Whole-Repository Stack Migration?

Researchers have introduced SWE Refactor Bench, a new evaluation framework designed to test whether AI coding agents can autonomously handle complex, repository-wide software migrations. While current tools excel at isolated bug fixes, this benchmark measures their ability to manage long-horizon tasks involved in updating outdated codebases. This effort addresses a critical bottleneck in software engineering where technical debt often requires expensive, manual intervention.