ai: How to run evals for the model router

The post explains how to evaluate the Foundry model router using an open-source GitHub repo for router-aware eval pipelines. It covers running quality, cost, and latency evaluations.

• Source: Microsoft AI Foundry Blog