← Glossary
Glossary · Infrastructure
model router
Also: AI router, LLM router
A service that sends model requests across multiple labs or hosts through one API.
A model router gives an application one integration while it chooses among several models or execution providers. It can route for price, speed, availability, geography, or a policy chosen by the developer.
The tradeoff is another layer between the application and the model. Provider-specific features and small API differences can still leak through the common interface.