GPT-5 in Azure AI Foundry featuring model router

Key insights

• Azure AI Foundry is a unified platform with over 11,000 AI models, including both open-source and proprietary options, allowing easy evaluation, fine-tuning, and secure deployment for enterprise needs.
  
• OpenAI’s GPT-OSS models are now available in Azure AI Foundry, supporting cloud and edge deployments through Foundry Local, which lets organizations run models on-premises or in the cloud while keeping data secure.
  
• The platform supports a flexible multi-model routing approach. This means users can select and route tasks to the best model for each situation, optimizing both performance and cost even though "model router" is not officially named as a product.
  
• GPT-4.5 preview, improved GPT-4 versions, and new features like the Azure OpenAI Assistants API (public preview) are now available. These offer advanced tools such as code interpreters, function calling, and expanded fine-tuning options.
  
• Additional enhancements include the Deep Research Agent for automating complex web research, the GPT-image-1 model for image generation/editing, pay-as-you-go compute options, improved security features like Entra Agent ID, and integration with Microsoft Purview for data governance.
  
• No direct release called "GPT-5 in Azure AI Foundry featuring model router" exists yet. However, organizations can benefit from continuous updates: latest GPT variants, hybrid deployment options across cloud/edge environments, implicit multi-model selection strategies, new agent-building tools, and strong focus on secure scalable infrastructure.
  

Keywords

GPT-5 Azure AI Foundry model router GPT-5 AI integration Azure AI models GPT-5 deployment Azure AI services model routing GPT-5 features in Azure