Farama-Foundation/chatarena
A Python library providing multi-agent language game environments for training and studying autonomous LLM agents.

Velocity · 7d
+1.3
★ / day
Trend
→steady
star history
ChatArena is a multi-agent simulation framework that provides language game environments for Large Language Models like GPT-3, GPT-4, and ChatGPT. It offers a flexible framework based on Markov Decision Process to define multiple players, environments, and their interactions. The library facilitates research on autonomous LLM agents and their social dynamics, supporting the development of communication and collaboration capabilities in AI systems.