RAGEN

mll-lab-nu/RAGEN is an open-source project by mll-lab-nu: Agent RL framework for LLM agents: multi-turn reinforcement learning with StarPO and reasoning-collapse diagnostics. Indexed here so it can be found — not resold.

It is free. Get it from the upstream repository: https://github.com/mll-lab-nu/RAGEN

From the project's own README (excerpt, reproduced for discovery under its MIT license):

RAGEN: Training Agents by Reinforcing Reasoning Diagnose agent failure modes. Make your RL training better.

RAGEN ( R easoning AGEN T) is a flexible RL framework for training reasoning LLM agents with multi-turn reinforcement learning , built on the StarPO algorithm. We develop diagnostics to understand how agentic reinforcement learning (agent RL) training works , and how to fix hidden failure modes such as the Echo Trap (RAGEN) and reasoning collapse (RAGEN-2).

-->

Project Page · RAGEN-2 Page · RAGEN on arXiv · RAGEN-2 on arXiv · MLL Lab (Manling Li, Northwestern)

> Looking for the V1 README? Please take a look here.

News • 2026.3.12. We are excited to release RAGEN-2 ! We introduce a systematic study of reasoning collapse in agent RL and lightweight interventions for stable training. See the v2 paper . • 2025.4.20. RAGEN V1 paper published on arXiv. • 2025.1.27. Initial RAGEN release. Post.

About

RAGEN is built around StarPO (State-Thinking-Actions-Reward Policy Optimization), a unified RL framework for training multi-turn, trajectory-level agents with flexible control over reasoning processes, reward assignment mechanisms, and prompt-rollout structures.

RAGEN is flexible with: • StarPO framework. Unified optimization for multi-turn agents, supporting both trajectory-level and turn-wise training. • 10 built-in environments. Sokoban, FrozenLake, WebShop, DeepCoder, SearchQA, Lean, Bandit, Countdown, MetaMathQA, Sudoku. • Gym-compatible interface. Easy to add custom environments.

RAGEN-2 additionally introduces: • SNR-Adaptive Filtering ( V2 ). Lightweight rollout filtering based on reward variance to mitigate noisy gradient updates. • Reasoning collapse diagnostics ( V2 ). Mutual information proxy metrics to detect…