Agent Lightning v1.0: Towards Harnessed Agentic RL

Hugging Face Daily Papers Papers
agentic-reinforcement-learning coding-agents reinforcement-learning framework reproducibility swe-bench

Summary

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.

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 arbitrary agents to RL training through an LLM endpoint proxy, an approach later adopted by frameworks such as verl Uni-Agent, AReaL 2.0, slime, and Polar. We refer to this paradigm as harnessed agentic RL, where the deploy-time harness directly participates in model post-training. Harnessed agentic RL differs fundamentally from traditional agentic RL: the harness, rather than the training engine, owns the environment interaction loop, while the trainer observes only sequences of LLM request-response pairs. This introduces challenges in retokenization, sample merging, advantage calculation, loss normalization, and backend scheduling, which can substantially affect training stability and effectiveness. We present Agent Lightning v1.0, a lightweight framework for harnessed agentic RL implemented in approximately 3,500 lines of code. It supports arbitrary agent harnesses and serves as a practical testbed for studying these challenges. We evaluate it on instruction-following, search, and coding agents, and provide a complete reproducible pipeline for coding-agent RL. Using only 6K training examples and modest compute, RL improves Qwen3.5-9B on SWE-bench Verified from 41.8% to 56.4%, a 14.6-point absolute gain. We release the complete workflow and training scripts to facilitate reproducible research on harnessed agentic RL.
Original Article
View Cached Full Text

Cached at: 08/19/26, 07:57 AM

Paper page - Agent Lightning v1.0: Towards Harnessed Agentic RL

Source: https://huggingface.co/papers/2608.17528

Abstract

Agent Lightning v1.0 enables reproducible reinforcement learning for arbitrary agent harnesses, substantially improving coding-agent performance with minimal data and compute.

Modern agents operate insideagent harnessesthat 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 arbitrary agents to RL training through anLLM endpoint proxy, an approach later adopted by frameworks such as verl Uni-Agent, AReaL 2.0, slime, and Polar. We refer to this paradigm asharnessed agentic RL, where the deploy-time harness directly participates in model post-training.Harnessed agentic RLdiffers fundamentally from traditional agentic RL: the harness, rather than the training engine, owns the environment interaction loop, while the trainer observes only sequences of LLM request-response pairs. This introduces challenges inretokenization, sample merging,advantage calculation,loss normalization, and backend scheduling, which can substantially affect training stability and effectiveness. We presentAgent Lightning v1.0, a lightweight framework forharnessed agentic RLimplemented in approximately 3,500 lines of code. It supports arbitraryagent harnessesand serves as a practical testbed for studying these challenges. We evaluate it on instruction-following, search, and coding agents, and provide a complete reproducible pipeline for coding-agent RL. Using only 6K training examples and modest compute, RL improves Qwen3.5-9B onSWE-bench Verifiedfrom 41.8% to 56.4%, a 14.6-point absolute gain. We release the complete workflow and training scripts to facilitate reproducible research onharnessed agentic RL.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2608\.17528

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/2608.17528 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2608.17528 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.17528 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

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.

ClawGym II: Exploring Black-Box RL on Agent Harness

Hugging Face Daily Papers

This paper introduces a unified black-box reinforcement learning framework for stable and scalable optimization of agents through complex harnesses, using sandbox execution and trajectory reconstruction with improvements on benchmarks like ClawGym-Bench.