One internal endpoint in front of all model providers that handles routing, keys, quotas, logging, caching and fallbacks.
You have done this if
Teams called your gateway instead of OpenAI or Azure directly, and you could see spend per team.
Say it in a review
All model traffic goes through the gateway, which gives us one place for keys, cost and fallbacks.
On the AI Application map LLM Gateway
Read Every model call should go through something you own · What an AI gateway actually costs to run