One model deployment to route them all
Microsoft Foundry's model router auto-selects the best AI model across 27 options from multiple providers
“what if I told you you can have a model router that selects the best model for you?”
Microsoft Foundry has launched a model router that can be deployed like any other model and automatically selects from 27 models across OpenAI, Claude, Grok, Llama, and others based on quality, cost, or balanced routing modes. The router dynamically picks the right model per task type — for example, routing math questions to a reasoning model like Grok 4.1 and simple tasks to a faster, cheaper model. This signals Microsoft's infrastructure bet on model-agnostic orchestration as a first-class deployment primitive.