AI RESEARCH
MonoScale: Scaling Multi-Agent System with Monotonic Improvement
arXiv CS.AI
•
ArXi:2601.23219v2 Announce Type: replace-cross In recent years, LLM-based multi-agent systems (MAS) have advanced rapidly, using a router to decompose tasks and delegate subtasks to specialized agents. A natural way to expand capability is to scale up the agent pool by continually integrating new functional agents or tool interfaces, but naive expansion can trigger performance collapse when the router cold-starts on newly added, heterogeneous, and unreliable agents.