Microsoft AI launches seven MAI models and Mayo Clinic partnership

The new model family covers reasoning, coding, image generation, voice, and transcription, with Microsoft Frontier Tuning designed to let organizations adapt models to their own workflows.

Microsoft AI has launched seven new MAI models covering reasoning, coding, image generation, voice, and transcription. Photo Credit: Mustafa Suleyman.

Microsoft AI has launched seven new in-house MAI models across reasoning, coding, image generation, voice, and transcription, alongside a healthcare AI collaboration with Mayo Clinic.

The announcement, made on 2 June 2026, includes MAI-Thinking-1, MAI-Code-1-Flash, MAI-Image-2.5, MAI-Image-2.5 Flash, MAI Transcribe-1.5, MAI-Voice-2, and MAI-Voice-2-Flash.

Microsoft AI says the models form a multimodal family designed to work across real-world tasks and integrate into products used by individuals, developers, and organizations. The launch also introduces Microsoft Frontier Tuning, an approach for adapting models to organization-specific workflows.

For developers and education providers focused on AI skills, the launch brings new model capabilities into coding, software engineering, voice, transcription, and enterprise workflow training. MAI-Code-1-Flash has been built for GitHub Copilot, Visual Studio Code, and Microsoft’s wider stack.

Microsoft AI is also working with Mayo Clinic to co-create a frontier AI model for healthcare, using Mayo Clinic’s clinical expertise, de-identified clinical data, and longitudinal insights alongside Microsoft’s foundational AI capabilities.

Seven models added to MAI family

MAI-Thinking-1 is Microsoft AI’s flagship reasoning model. Microsoft AI says the model was trained from the ground up on clean data, without distillation from third-party models, and matches leading models on key software engineering benchmarks.

Mustafa Suleyman, CEO of Microsoft AI, said on LinkedIn that MAI-Thinking-1 is a "35B active parameter MoE with a 256K context window" and had reached "97% on AIME 2025" and "53% on SWE Bench Pro."

MAI-Code-1-Flash is an inference-efficient agentic coding model designed for GitHub Copilot, Visual Studio Code, and the Microsoft stack. Suleyman said the model has five billion parameters and is "especially tuned for VS Code and GitHub Copilot CLI."

Microsoft AI says MAI-Image-2.5 and MAI-Image-2.5 Flash support text-to-image generation and image editing. MAI Transcribe-1.5 is positioned for transcription across 43 languages, while MAI-Voice-2 supports speech generation across 15 languages and can adapt to a voice from a short sample with safeguards against misuse.

MAI-Voice-2-Flash is listed as coming soon, with Microsoft AI describing it as a lower-cost and more efficient package.

Frontier Tuning targets organization-specific AI

Microsoft Frontier Tuning is designed to let organizations adapt MAI models to their own workflows using reinforcement learning environments.

Microsoft AI describes those environments as training spaces where models can learn from the sequence of work, decisions, and actions inside an organization. The company says customer data remains within the organization’s environment and control.

In his LinkedIn post, Suleyman said Frontier Tuning lets users "customize our models to create custom, company-specific agents that only you control."

He added: "You can make our model, your model. Your data. Your agents. Your moat."

Microsoft AI says early tuning work has shown efficiency and performance gains. The announcement says an MAI tuned model for Excel matched GPT 5.4 while being up to 10 times more efficient, while a model tuned for McKinsey achieved the highest win rate of any model tested at around 10 times lower cost.

Suleyman said on LinkedIn that, when tuned for McKinsey’s tasks, MAI "outperforming GPT-5.5 on quality, while being 10x lower on cost."

Mayo Clinic model planned for healthcare

Microsoft and Mayo Clinic are collaborating on a frontier AI model for healthcare that will be designed for clinical reasoning and healthcare use cases.

Microsoft AI says the model will first be deployed within Mayo Clinic’s own environment. Once validated, the model is expected to be made available to other organizations through Azure Foundry.

The model will be owned by Mayo Clinic, with Microsoft AI pointing to patient trust, clinical rigor, safety, and stewardship of clinical health data and AI.

The launch also adds detail to Microsoft AI’s broader strategy. Microsoft AI says it is building what it calls a superintelligence lab, with in-house models trained from scratch, clean and appropriately licensed datasets, technical and safety reports, and co-design work with Maia 200 silicon.

Microsoft AI says the next year will bring a rapid scale-up in compute and capabilities. The MAI models are available now, with MAI-Voice-2-Flash listed as coming soon and the Mayo Clinic healthcare model moving first into Mayo Clinic’s own environment before wider availability through Azure Foundry.