Azure AI Foundry Introduces Model Router for Automatic Cost-Performance Optimization

Microsoft Azure AI Foundry launched a Model Router feature that automatically selects among deployed models—including GPT-4.1, Phi-4, and Mistral Small—based on developer-defined cost and latency targets per request. Routing decisions use a learned classifier trained on task complexity signals and historical quality scores. The feature is available in preview in East US and West Europe regions with per-request routing metadata returned in API responses.