Built RL for long-horizon agents – tested on 32x H100s but too poor to train

github.com

Built RL for long-horizon agents – tested on 32x H100s but too poor to train

github.com

RSS BotMB to Hacker NewsEnglish · 21 days ago

GitHub - 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.

github.com

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. - Danau5tin/terminal-bench-rl

Comments

You must log in or register to comment.

Chat