Microsoft is expanding its model options by adding OpenAI GPT-image-1-mini, GPT-realtime-mini, and GPT-audio-mini, and the upgraded GPT-5, to its developer toolkit. In the words of a Microsoft press release:
By expanding Azure AI Foundry with the latest OpenAI models and advancing our agentic AI framework, we empower customers with unparalleled choice, flexibility, and business capabilities. This enables developers to build intelligent agent systems that address complex business needs and drive innovation at scale.
So, what are these models, and how do they enhance the existing offerings in Azure AI Foundry?
AI Agent & Copilot Summit is an AI-first event to define opportunities, impact, and outcomes with Microsoft Copilot and agents. Building on its 2025 success, the 2026 event takes place March 17-19 in San Diego. Get more details.
Multimodal Expansion
Microsoft describes this multimodal expansion as enabling developers to “unlock the full spectrum of AI: text, images, audio, and video.” Collectively, the introduction of these models does just that.
- GPT-image-1-mini — GPT-image-1-mini is designed to enable very high-quality text-to-image and image-to-image capabilities while not consuming vast computational resources. It’s quick, efficient, and scalable, allowing real-time image generation integrated with Azure AI Foundry workflows.
- GPT-realtime-mini — This fast, lightweight model is designed for real-time voice interaction. It is cost-effective, powerful, and ideally suited for speed and responsiveness with very low-latency.
- GPT-audio-mini — GPT-audio-mini is a streamlined audio model developed for dynamic audio generation, suitable for tasks such as voiceovers. It supports the mission to provide multimodal capabilities at a reduced cost.
- GPT-5-chat-latest — Now updated in Azure AI Foundry, the latest version of the model includes improved safety and guardrails. In particular, the model is equipped to more effectively recognize and manage outputs that could lead to mental or emotional distress.
- GPT-5-pro — GPT-5-pro offers what Microsoft refers to as research-grade intelligence, positioning it at the forefront of analysis and reasoning capabilities. Deploying this model through Azure AI Foundry allows developers to maximize its potential by leveraging multiple reasoning pathways to deliver the most accurate and reliable results.
Closing Thoughts
A few solid facts about Azure AI Foundry. The catalog features an extensive list of over 1,900 models. They are capable of a comprehensive range of tasks and include various types of models, from foundation models to domain-specific and industry models.
As well as OpenAI, and in-house options, model partners include massive names including Meta, Mistral AI, DeepSeek, xAI, Hugging Face, NVIDIA, and Cohere. I wanted to provide this context to give you a scope on the scale of the platform and what was already on offer to developers.
What’s different now, as Microsoft puts it:
Developers can now move beyond text, tapping into image and audio generation, editing, and understanding.
This is big news. However, if we delve a little deeper, it’s clear that despite reports suggesting a cooling relationship between OpenAI and Microsoft in recent months, Sam Altman has been steering OpenAI into new directions with major partners —most notably Oracle — and large infrastructure projects like Stargate, the relationship between OpenAI and Microsoft continues to not only exist, but, in places, grow.
This reality is crucial for Microsoft as a company, as well as for its customers. Azure AI Foundry faces fierce competition — particularly from Amazon Bedrock and Google Cloud’s Vertex AI. However, by continuing to adopt the latest models at a rapid pace and maintaining its partnership with OpenAI (despite clear signs of a growing split), Microsoft has positioned Azure AI Foundry as a strong contender to become the leading platform for AI and agent development.
Ask Cloud Wars AI Agent about this analysis
