AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xi, Zhiheng, Huang, Jixuan, Liao, Chenyang, Huang, Baodai, Guo, Honglin, Liu, Jiaqi, Zheng, Rui, Ye, Junjie, Zhang, Jiazheng, Chen, Wenxiang, He, Wei, Ding, Yiwen, Li, Guanyu, Chen, Zehui, Du, Zhengyin, Yao, Xuesong, Xu, Yufei, Chen, Jiecao, Gui, Tao, Wu, Zuxuan, Zhang, Qi, Huang, Xuanjing, Jiang, Yu-Gang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911148006178816
author Xi, Zhiheng
Huang, Jixuan
Liao, Chenyang
Huang, Baodai
Guo, Honglin
Liu, Jiaqi
Zheng, Rui
Ye, Junjie
Zhang, Jiazheng
Chen, Wenxiang
He, Wei
Ding, Yiwen
Li, Guanyu
Chen, Zehui
Du, Zhengyin
Yao, Xuesong
Xu, Yufei
Chen, Jiecao
Gui, Tao
Wu, Zuxuan
Zhang, Qi
Huang, Xuanjing
Jiang, Yu-Gang
author_facet Xi, Zhiheng
Huang, Jixuan
Liao, Chenyang
Huang, Baodai
Guo, Honglin
Liu, Jiaqi
Zheng, Rui
Ye, Junjie
Zhang, Jiazheng
Chen, Wenxiang
He, Wei
Ding, Yiwen
Li, Guanyu
Chen, Zehui
Du, Zhengyin
Yao, Xuesong
Xu, Yufei
Chen, Jiecao
Gui, Tao
Wu, Zuxuan
Zhang, Qi
Huang, Xuanjing
Jiang, Yu-Gang
contents Developing autonomous LLM agents capable of making a series of intelligent decisions to solve complex, real-world tasks is a fast-evolving frontier. Like human cognitive development, agents are expected to acquire knowledge and skills through exploration and interaction with the environment. Despite advances, the community still lacks a unified, interactive reinforcement learning (RL) framework that can effectively train such agents from scratch -- without relying on supervised fine-tuning (SFT) -- across diverse and realistic environments. To bridge this gap, we introduce AgentGym-RL, a new framework to train LLM agents for multi-turn interactive decision-making through RL. The framework features a modular and decoupled architecture, ensuring high flexibility and extensibility. It encompasses a wide variety of real-world scenarios, and supports mainstream RL algorithms. Furthermore, we propose ScalingInter-RL, a training approach designed for exploration-exploitation balance and stable RL optimization. In early stages, it emphasizes exploitation by restricting the number of interactions, and gradually shifts towards exploration with larger horizons to encourage diverse problem-solving strategies. In this way, the agent develops more diverse behaviors and is less prone to collapse under long horizons. We perform extensive experiments to validate the stability and effectiveness of both the AgentGym-RL framework and the ScalingInter-RL approach. Our agents match or surpass commercial models on 27 tasks across diverse environments. We offer key insights and will open-source the complete AgentGym-RL framework -- including code and datasets -- to empower the research community in developing the next generation of intelligent agents.
format Preprint
id arxiv_https___arxiv_org_abs_2509_08755
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning
Xi, Zhiheng
Huang, Jixuan
Liao, Chenyang
Huang, Baodai
Guo, Honglin
Liu, Jiaqi
Zheng, Rui
Ye, Junjie
Zhang, Jiazheng
Chen, Wenxiang
He, Wei
Ding, Yiwen
Li, Guanyu
Chen, Zehui
Du, Zhengyin
Yao, Xuesong
Xu, Yufei
Chen, Jiecao
Gui, Tao
Wu, Zuxuan
Zhang, Qi
Huang, Xuanjing
Jiang, Yu-Gang
Machine Learning
Artificial Intelligence
Computation and Language
Developing autonomous LLM agents capable of making a series of intelligent decisions to solve complex, real-world tasks is a fast-evolving frontier. Like human cognitive development, agents are expected to acquire knowledge and skills through exploration and interaction with the environment. Despite advances, the community still lacks a unified, interactive reinforcement learning (RL) framework that can effectively train such agents from scratch -- without relying on supervised fine-tuning (SFT) -- across diverse and realistic environments. To bridge this gap, we introduce AgentGym-RL, a new framework to train LLM agents for multi-turn interactive decision-making through RL. The framework features a modular and decoupled architecture, ensuring high flexibility and extensibility. It encompasses a wide variety of real-world scenarios, and supports mainstream RL algorithms. Furthermore, we propose ScalingInter-RL, a training approach designed for exploration-exploitation balance and stable RL optimization. In early stages, it emphasizes exploitation by restricting the number of interactions, and gradually shifts towards exploration with larger horizons to encourage diverse problem-solving strategies. In this way, the agent develops more diverse behaviors and is less prone to collapse under long horizons. We perform extensive experiments to validate the stability and effectiveness of both the AgentGym-RL framework and the ScalingInter-RL approach. Our agents match or surpass commercial models on 27 tasks across diverse environments. We offer key insights and will open-source the complete AgentGym-RL framework -- including code and datasets -- to empower the research community in developing the next generation of intelligent agents.
title AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2509.08755