AI RESEARCH
Notation Matters: A Benchmark Study of Token-Optimized Formats in Agentic AI Systems
arXiv CS.AI
•
ArXi:2605.29676v1 Announce Type: new Large language models in Agentic AI systems consume tool schemas and execution results and emit tool invocations as structured data. The default language for that exchange, JSON, was designed for application-to-application interchange rather than token efficiency, so its structural elements impose substantial token overhead.