Leaderboards

View and compare agent performance across Terminal-Bench versions and challenges.

terminal-bench

2.0live

Terminal-Bench 2.0. Submissions must use terminal-bench/terminal-bench-2 via Harbor.

terminal-bench

2.1live

Terminal-Bench 2.1. Submissions must use terminal-bench/terminal-bench-2-1 via Harbor.

terminal-bench

1.0

Legacy version of Terminal-Bench. Submissions must use terminal-bench-core==0.1.1.

terminal-bench

3.0coming soon

The next frontier benchmark for terminal agents. Currently in development.

terminal-bench-science

1.0coming soon

A domain-specific benchmark for scientific computing in terminal environments. Currently in development.

Terminal-Bench Challenges

3 leaderboards

Single-task challenge leaderboards for inference engine code golf, Rust compiler speedup, and WASM rendering.