MetaGen: Self-Evolving Roles and Topologies for Multi-Agent LLM Reasoning

TL;DR

MetaGen is a training-free framework that dynamically adapts roles and interaction topologies in multi-agent LLM systems during inference, enhancing accuracy and efficiency without updating model weights.

Contribution

It introduces a novel method for real-time adaptation of roles and collaboration structures in multi-agent LLMs, overcoming rigidity of fixed role libraries and static topologies.

Findings

Improves accuracy on code generation and reasoning benchmarks.

Reduces inference cost compared to fixed-topology baselines.

Demonstrates effective dynamic role and topology adaptation during inference.

Abstract

Large language models are increasingly deployed as multi-agent systems, where specialized roles communicate and collaborate through structured interactions to solve complex tasks that often exceed the capacity of a single agent. However, most existing systems still rely on a fixed role library and an execution-frozen interaction topology, a rigid design choice that frequently leads to task mismatch, prevents timely adaptation when new evidence emerges during reasoning, and further inflates inference cost. We introduce MetaGen, a training-free framework that adapts both the role space and the collaboration topology at inference time, without updating base model weights. MetaGen generates and rewrites query-conditioned role specifications to maintain a controllable dynamic role pool, then instantiates a constrained execution graph around a minimal backbone. During execution, it iteratively updates role prompts and adjusts structural decisions using lightweight feedback signals. Experiments on code generation and multi-step reasoning benchmarks show that MetaGen improves the accuracy and cost tradeoff over strong multi-agent baselines.