awesome-multi-agent-scaling

A curated list of papers related to scaling laws for multi-agent systems (MAS).

Performance

• More agents is all you need (Paper, Code)
• Scaling Large-Language-Model-based Multi-Agent Collaboration ([Paper](Scaling Large-Language-Model-based Multi-Agent Collaboration), Code)
• The Stability, Scalability and Performance of Multi-agent Systems (Paper)

Heterogeneity

• Mixture of Agents (Paper, Code)
• Internet of Agents (Paper, Code)
• Diversity of Thought Elicits Stronger Reasoning Capabilities in Multi-Agent Debate Frameworks (Paper)
• Rethinking Mixture-of-Agents: Is Mixing Different Large Language Models Beneficial? (Paper)
• CoMM: Collaborative Multi-Agent, Multi-Reasoning-Path Prompting for Complex Problem Solving (Paper)
• X-MAS: Towards Building Multi-Agent Systems with Heterogeneous LLMs (Paper)

Multi-agent Reasoning

• Improving LLM Reasoning with Multi-Agent Tree-of-Thought Validator Agent (Paper)
• MALT: Improving Reasoning with Multi-Agent LLM Training (Paper)
• Exchange-of-Thought: Enhancing Large Language Model Capabilities through Cross-Model Communication (Paper)

Automated Agent Generation

• EvoAgent: Towards Automatic Multi-Agent Generation via Evolutionary Algorithms (Paper)
• Automated Design of Agentic Systems (Paper)
• Dynamic LLM-Agent Network: An LLM-agent Collaboration Framework with Agent Team Optimization (Paper)
• Gptswarm: Language agents as optimizable graphs (Paper)
• Agentverse: Facilitating multi-agent collaboration and exploring emergent behaviors (Paper)
• Agent Symbolic Learning: Symbolic learning enables self-evolving agents (Paper)
• Multi-agent Architecture Search via Agentic Supernet (Paper)
• AgentNet: Decentralized Evolutionary Coordination for LLM-based Multi-Agent Systems (Paper)

Infrastructure

• AgentScope: A Flexible yet Robust Multi-Agent Platform (Paper, Code)
• Very Large-Scale Multi-Agent Simulation in AgentScope (Paper)
• A Scalable Communication Protocol for Networks of Large Language Models (Paper)
• AI Metropolis: Scaling Large Language Model-based Multi-Agent Simulation with Out-of-order Execution (Paper)
• DAWN: Designing Distributed Agents in a Worldwide Network (Paper)