SWE-bench Pro

'SWE-bench Pro' is a benchmark dataset used to evaluate the performance of coding agents in software engineering tasks, specifically focusing on their ability to explore code repositories efficiently.

Papers using SWE-bench Pro (3)

FastContext: Training Efficient Repository Explorer for Coding Agents2026SkillsVote: Lifecycle Governance of Agent Skills from Collection, Recommendation to Evolution2026Socratic-SWE: Self-Evolving Coding Agents via Trace-Derived Agent Skills2026