Pinned Loading

1. terminal-bench-rl terminal-bench-rl Public

   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.

   Python 230 14

2. calculator_agent_rl calculator_agent_rl Public

   Training an LLM to use a calculator with multi-turn reinforcement learning, achieving a **62% absolute increase in evaluation accuracy**.

   Python 45 3

3. tbench-agentic-data-pipeline tbench-agentic-data-pipeline Public

   Multi-agent synthetic data generation pipeline capable of generating and validating long horizon terminal/coding tasks for RL training

   Python 13 2

4. agentic_environments agentic_environments Public

   A mini-framework to build agentic LLM environments.

   Python 2