Microsoft Introduces Seven New MAI Models and Announces Healthcare Collaboration with Mayo Clinic

Microsoft announced in a press release the launch of seven new models under its MAI family, built by its in-house AI division. The lineup includes MAI-Thinking-1 for reasoning, MAI-Code-1-Flash for coding tasks, MAI-Image-2.5 for image generation and editing, MAI-Transcribe-1.5 for multilingual transcription, and MAI-Voice-2 for natural speech generation. Each model is designed to work across Microsoft products and will also be available to developers through OpenRouter, Fireworks, and Baseten.

The company introduced a new reinforcement learning approach called Frontier Tuning, which allows organizations to adapt MAI models using their own data and workflows. This approach enables models to learn from real-world processes while keeping proprietary data under customer control. Microsoft reported that internally tuned MAI models have achieved efficiency gains of up to ten times compared to general-purpose systems.

Microsoft also revealed a partnership with Mayo Clinic to co-develop a frontier AI model for healthcare. The model will combine Mayo Clinic’s clinical data and expertise with Microsoft’s AI capabilities to support clinical reasoning and treatment planning. It will first be deployed within Mayo Clinic’s environment before being made available to other healthcare organizations through Microsoft Foundry.

The announcement marks part of Microsoft AI’s broader initiative to build what it calls a “hill-climbing machine,” a long-term research effort focused on continuous improvement through clean data, proprietary infrastructure, and transparent evaluation practices.