Augment Code debuts AI agent with 70% win rate over GitHub Copilot and record-breaking SWE-bench score
1 min read
Summary
Augment Code has launched Augment Agent, an AI tool which helps developers navigate and modify large codebases, rather than just generating simple code.
The tool enables a real-time synchronisation of code changes across developer teams, and a “Memories” function, which learns from user interactions to align with individual coding styles and preferences.
Augment Agent has achieved the highest score to date on SWE-bench verified, an industry benchmark for AI coding capabilities, through its use of Anthropic’s Claude Sonnet 3.7 and OpenAI’s O1 reasoning model.
“The challenge for any AI system, including Augment, is that when you’re working with large systems containing tens of millions of lines of code – which is typical for meaningful software applications – you simply can’t pass all that as context to today’s large language models,” said Scott Dietzen, CEO of Augment Code.