Microsoft Announces 3 New MAI Models Available on Foundry

Microsoft has unveiled 3 new MAI models immediately available on Foundry and MAI Playground. These new models feature excellent quality, fast speed, and competitive pricing.

MAI-Transcribe-1 delivers state-of-the-art speech-to-text (STT) capabilities for the top 25 languages. It offers 2.5x faster batch transcription than existing Microsoft Azure Fast, with reasonable pricing starting at $0.36 per hour.

MAI-Voice-1 is a top-tier voice generation model that produces natural, emotionally rich speech. It can safely generate a custom voice from just a few seconds of audio, boasting overwhelming speed that can generate 60 seconds of audio in just 1 second.

MAI-Image-2 significantly enhances image generation performance and speed. It offers at least 2x faster generation speed compared to existing models on Foundry and Copilot, featuring natural lighting and skin tones along with accurate text rendering. Pricing starts at $5 per 1 million text input tokens and $33 per 1 million image output tokens.

Microsoft AI aims for Humanist AI that puts humans at the center and is optimized for how people actually communicate, and is rapidly deploying these models across its products and commercial services.