Running benchmarks

To run benchmark tasks with the minimal agent composition, follow Get started with the Python SDK. The guide covers installation, running minimal.cordis.yml, and isolating workspaces and session IDs between tasks.