Cost-Aware Model Orchestration for LLM-based Systems

ArXi:2512.01099v2 Announce Type: replace As modern artificial intelligence (AI) systems become advanced and capable, they can leverage a wide range of tools and models to perform complex tasks. The task of orchestrating these models is increasingly performed by Large Language Models (LLMs) that rely on qualitative descriptions of models for decision-making. However, the descriptions provided to existing LLM-based orchestrators frequently do not reflect true model capabilities and performance characteristics, leading to suboptimal model selection, reduced task accuracy, and increased cost.