Agent Lightning: Train ANY AI Agents with Reinforcement Learning

Papers with Code Trending Papers
reinforcement-learning llm ai-agents framework hierarchical-rl training-framework

Summary

Agent Lightning introduces a flexible reinforcement learning framework for training large language models in AI agents, achieving decoupling between agent execution and training to handle complex interactions.

We present Agent Lightning, a flexible and extensible framework that enables Reinforcement Learning (RL)-based training of Large Language Models (LLMs) for any AI agent. Unlike existing methods that tightly couple RL training with agent or rely on sequence concatenation with masking, Agent Lightning achieves complete decoupling between agent execution and training, allowing seamless integration with existing agents developed via diverse ways (e.g., using frameworks like LangChain, OpenAI Agents SDK, AutoGen, and building from scratch) with almost ZERO code modifications. By formulating agent execution as Markov decision process, we define an unified data interface and propose a hierarchical RL algorithm, LightningRL, which contains a credit assignment module, allowing us to decompose trajectories generated by ANY agents into training transition. This enables RL to handle complex interaction logic, such as multi-agent scenarios and dynamic workflows. For the system design, we introduce a Training-Agent Disaggregation architecture, and brings agent observability frameworks into agent runtime, providing a standardized agent finetuning interface. Experiments across text-to-SQL, retrieval-augmented generation, and math tool-use tasks demonstrate stable, continuous improvements, showcasing the framework's potential for real-world agent training and deployment.
Original Article
View Cached Full Text

Cached at: 08/25/26, 04:31 AM

Paper page - Agent Lightning: Train ANY AI Agents with Reinforcement Learning

Source: https://huggingface.co/papers/2508.03680

Abstract

Agent Lightning is a flexible RL framework for training LLMs in various agents, using a hierarchical RL algorithm and decoupling execution from training to handle complex interactions.

We present Agent Lightning, a flexible and extensible framework that enablesReinforcement Learning(RL)-based training ofLarge Language Models(LLMs) for any AI agent. Unlike existing methods that tightly couple RL training with agent or rely on sequence concatenation with masking, Agent Lightning achieves complete decoupling between agent execution and training, allowing seamless integration with existing agents developed via diverse ways (e.g., using frameworks like LangChain, OpenAI Agents SDK, AutoGen, and building from scratch) with almost ZERO code modifications. By formulating agent execution asMarkov decision process, we define an unified data interface and propose ahierarchical RL algorithm, LightningRL, which contains a credit assignment module, allowing us to decompose trajectories generated by ANY agents into training transition. This enables RL to handle complex interaction logic, such as multi-agent scenarios and dynamic workflows. For the system design, we introduce aTraining-Agent Disaggregation architecture, and brings agent observability frameworks into agent runtime, providing a standardized agent finetuning interface. Experiments acrosstext-to-SQL, retrieval-augmented generation, andmath tool-use tasksdemonstrate stable, continuous improvements, showcasing the framework’s potential for real-world agent training and deployment.

View arXiv pageView PDFProject pageGitHub17.7kAdd to collection

Get this paper in your agent:

hf papers read 2508\.03680

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2508.03680 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2508.03680 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2508.03680 in a Space README.md to link it from this page.

Collections including this paper56

Browse 56 collections that include this paper

Similar Articles

Agent Lightning v1.0

Hacker News Top

Agent Lightning v1.0 is a lightweight agentic reinforcement learning framework by Microsoft, refactored for training AI agents with real harnesses and achieving substantial benchmark improvements, such as a 14.6 percentage point gain on SWE-bench.

Agent Lightning v1.0: Towards Harnessed Agentic RL

Hugging Face Daily Papers

Agent Lightning v1.0 is a lightweight framework that enables reproducible reinforcement learning for agent harnesses, significantly boosting coding-agent performance on benchmarks like SWE-bench Verified.

AgentJet: A Flexible Swarm Training Framework for Agentic Reinforcement Learning

arXiv cs.AI

AgentJet is a distributed swarm training framework for LLM agent reinforcement learning that decouples agent rollouts from model optimization, enabling heterogeneous multi-agent RL, multi-task training, fault tolerance, and live code iteration with 1.5-10x training speedup. It also introduces an automated research system capable of autonomously conducting multi-day RL studies on large-scale clusters.