Microsoft launches three in-house AI models

Microsoft has released three in-house foundational AI models — MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 — targeting enterprise use cases in transcription, voice, and image generation. The move puts Microsoft in direct competition with enterprise AI rivals despite its close ties with OpenAI. MAI-Transcribe-1 supports 25 languages at 50% lower GPU costs than alternatives, while MAI-Voice-1 generates 60 seconds of audio in under one second. MAI-Image-2, built with artist collaboration, debuted third on Arena.ai's image leaderboard. All models are available via Microsoft Foundry and MAI Playground.