AgentGym-RL and AgentGym
These are successive versions of the same research project, with AgentGym-RL focusing specifically on multi-turn reinforcement learning for long-horizon tasks while the newer AgentGym broadens the scope to agent training across diverse environments.
About AgentGym-RL
WooooDyy/AgentGym-RL
Code and implementations for the paper "AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning" by Zhiheng Xi et al.
This framework helps developers train large language model (LLM) agents to make intelligent decisions over many steps in real-world scenarios. It takes an LLM and training data from diverse environments as input, and outputs an enhanced LLM agent capable of multi-turn interactions that can match or surpass commercial models. Machine learning researchers and practitioners focused on agent development would use this.
About AgentGym
WooooDyy/AgentGym
Code and implementations for the ACL 2025 paper "AgentGym: Evolving Large Language Model-based Agents across Diverse Environments" by Zhiheng Xi et al.
AgentGym is a framework that allows AI researchers to develop and evaluate large language model-based agents across a wide range of tasks and environments. It takes in an LLM agent and provides standardized feedback from diverse environments like web browsing, text games, and digital tasks. The output is an evaluated agent, its performance metrics, and detailed interaction trajectories, helping researchers understand and improve agent behaviors. This is for AI researchers and practitioners focused on building capable, generalist LLM agents.
Scores updated daily from GitHub, PyPI, and npm data. How scores work