Microsoft AI has released the MAI Models

Hi All,

Microsoft has introduced a new generation of in‑house AI models under the Microsoft AI (MAI) brand, marking a major step toward a more independent and multimodal AI platform

Models

• MAI-Voice-1: MAI-Voice-1 is our next-generation voice model, designed to produce natural, expressive speech that adapts to your intent.Learn more here
• MAI-Image-2: MAI‑Image‑2 is our most capable image model yet, built to deliver grounded, reliable visuals for real creative work. Learn more here
• MAI-1-Preview: MAI-1-Preview is our first end-to-end foundational model, offering an early look at the intelligence coming to everyone.Not currently available.
• MAI-Transcribe-1: Our most accurate speech-to-text model delivering fast, reliable transcription across 25 languages in real‑world conditions. Learn more here

Sadly the Playground does not work here

Azure Foundry

Let’s spin up an Azure Foundry Resource

Check the Models here

Open in Playground

Play around here

Regards
Andres Bohren