Definition

Model Routing

Model routing sends each query to the cheapest model that can handle it well. A cheap classifier call labels the request simple or complex; simple ones go to GPT-4o-mini or Claude Haiku, complex ones to a frontier model. Teams typically save 40 to 60 percent of LLM spend with no measurable quality loss, because simple queries never used the big model's capability. Reasoning models add a thinking-time axis.