Agent Lightning v1.0 Advances Harnessed Agentic RL

Zhiyuan He, Siwei Zhang, Zhiwen Zhou, Yuqing Yang, Yu Kang, Yuge Zhang, Luna K. Qiu, Tin Yan Tsui, Jiahang Xu, Chong Luo· August 19, 2026 View original

Key takeaways

  • Agent Lightning v1.0 is a framework for harnessed agentic RL.
  • It addresses challenges in training agents where the harness controls interaction.
  • The framework significantly improves coding agent performance (e.g., SWE-bench).
  • It is lightweight and provides a reproducible research pipeline.

Who benefits

Software DevelopmentAI ResearchRoboticsAutomationIT Services

Summary

Agent Lightning v1.0 is a new lightweight framework for harnessed agentic reinforcement learning (RL) that addresses challenges in training agents where the harness, not the training engine, controls environment interaction. It shows significant performance gains for coding agents.

Modern AI agents often operate within "agent harnesses" that manage their tools, context, and control flow, making the harness a crucial component of the overall system. This paper introduces Agent Lightning v1.0, a lightweight framework designed for "harnessed agentic reinforcement learning" (RL). This paradigm fundamentally differs from traditional agentic RL because the harness, rather than the training engine, directly handles the environment interaction loop. This introduces unique challenges in areas like retokenization, sample merging, and loss normalization, which can impact training stability and effectiveness. Agent Lightning v1.0, implemented in just 3,500 lines of code, provides a practical testbed to study these challenges. It supports arbitrary agent harnesses and demonstrates significant improvements, such as boosting a Qwen3.5-9B model's performance on SWE-bench Verified from 41.8% to 56.4% with only 6,000 training examples. The complete workflow and training scripts are released to foster reproducible research.

Why it matters

This framework provides a practical solution for training and improving complex AI agents that rely on external tools and environments, offering significant performance gains for tasks like coding and instruction-following.

How to implement this in your domain

  1. 1Download and experiment with Agent Lightning v1.0 for training custom AI agents.
  2. 2Adapt existing agent harnesses to integrate with the Agent Lightning framework for RL training.
  3. 3Utilize the provided reproducible pipeline for coding-agent RL to enhance internal development.
  4. 4Contribute to or leverage the open-source community around harnessed agentic RL for best practices.

Original post by Zhiyuan He, Siwei Zhang, Zhiwen Zhou, Yuqing Yang, Yu Kang, Yuge Zhang, Luna K. Qiu, Tin Yan Tsui, Jiahang Xu, Chong Luo

"arXiv:2608.17528v1 Announce Type: new Abstract: Modern agents operate inside agent harnesses that manage tools, context, and control flow, making the harness a critical part of the agent system. Our original Agent Lightning introduced a disaggregated architecture that connects ar…"

View on X

Originally posted by Zhiyuan He, Siwei Zhang, Zhiwen Zhou, Yuqing Yang, Yu Kang, Yuge Zhang, Luna K. Qiu, Tin Yan Tsui, Jiahang Xu, Chong Luo on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses

More in AI Engineering & DevTools