Large Language Models (LLMs) as planning and decision-makers have marked a new milestone in AI’s application in autonomous tasks. This survey presents a detailed examination of multi-agent systems utilizing LLMs.

• Highlights the domains and environments where LLM-based agents operate.
• Discusses agents’ profiling and communication methods.
• Indicates mechanisms that facilitate agent capacity growth.
• Provides access to datasets and benchmarks critical for researchers in this area.

Offering a 360-degree view of the field, this survey is instrumental for anyone interested in the collaborative and interactive dimensions of AI. The insights presented here can foster innovation and progress in creating more complex and effective multi-agent AI systems.