Benchmark Almanac

Augment Code

We track 2 benchmarks that Augment Code made or helped make.

0 in use 1 with public code 1 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
Augment Prism multi-turn coding benchmark Coding agents follow a queued conversation and edit a repository. A model compares the final changes with the original pull request. Fixing real repositories May 2026 0.21 Not in cards
AI Code Review Evaluation (Augment Code) Agents review pull requests from open source projects. Augment compares their comments with a corrected reference set. Fixing real repositories Dec 2025 Not in cards