Microsoft has officially unveiled its first in-house artificial intelligence models — MAI-Voice-1 and MAI-1-preview. MAI-Voice-1 is a high-fidelity speech generation model capable of synthesizing a minute of audio in less than a second on a single GPU, and is already integrated into Copilot Daily, podcasts, and available through Copilot Labs. MAI-1-preview is Microsofts first foundational language model, trained on ~15,000 NVIDIA H100 GPUs; it is undergoing public testing on the LMArena platform. This move underscores Microsofts push for AI autonomy and reduced reliance on third-party partners like OpenAI.
Microsoft Unveils First In-House AI Models: MAI-Voice-1 and MAI-1-preview
