Microsoft ships MAI-Transcribe-1, MAI-Voice-1, MAI-Image-2 in Azure AI Foundry

Microsoft released three production-ready first-party MAI models: Transcribe-1 (speech-to-text in 25 languages, 2.5x faster), Voice-1 (hyper-realistic branded voice generation), and Image-2 (campaign-ready image generation). All target enterprise marketing and creative workflows.