Software
A new system to test how software agents find and fix bugs
The system creates executable tasks from historical code changes to evaluate software agents.
The system creates executable tasks from historical code changes to evaluate software agents.
The Code Skill Server finds useful skills to improve how models generate code.
A framework called SWE-at-Scale uses historical bug fixes to build diverse and challenging tasks.
Researchers found that existing mutation tools and benchmarks fail to represent 86.3% of real-world quantum software bugs.