Danau5tin/terminal-bench-rl

GRPO training code which scales to 32xH100s for long horizon terminal/coding tasks. Base agent is now the top Qwen3 agent on Stanford's TerminalBench leaderboard.

125 best points · 12 comments · 1 Hacker News post · 419 GitHub stars · View on GitHub

Hacker News discussions

• Show HN: Terminal-Bench-RL: Training long-horizon terminal agents with RL (125 points, 12 comments, by Danau5tin, 2025-07-29)