MAI Foundation Models

Microsoft's first major foundation model push, releasing three MAI models in Azure Foundry — a direct shot at OpenAI and Google dependencies. MAI-Transcribe-1 (speech-to-text, 25 languages) ranks #1 by FLEURS in 11 core languages, beating Whisper-large-v3 and Gemini 3.1 Flash-Lite. MAI-Voice-1 (text-to-speech) generates 60 seconds of audio in under 1 second on a single GPU. MAI-Image-2 (text-to-image) debuted #3 on Arena.ai's image leaderboard.

Marks Microsoft's transition from pure OpenAI dependency to building its own frontier multimodal models internally under VP GenAI Mustafa Suleyman's MAI division. Proprietary, available via Azure Foundry.