Model Routing
An architecture pattern in LLM applications where a dynamic gateway analyzes incoming queries and routes them to the most appropriate, cost-effective model.
Think of It Like This
Like a telephone operator who connects a simple question to a junior clerk, but routes a complex legal issue to a senior partner.
By utilizing model routing, applications balance performance and cost. A cheap, fast model handles basic queries or summarization, while a massive reasoning model is reserved for complex logical tasks. The routing logic itself is often governed by a smaller machine learning classifier or heuristic rules.