Microsoft AI has launched its two in-house AI models, MAI-Voice-1 and MAI-1-preview, where "MAI" stands for Microsoft AI.
The American tech giant stated that its new MAI-Voice-1 speech model can create a minute’s worth of audio in under one second on just one GPU, while MAI-1-preview, providing a “glimpse of future offerings inside Copilot.”
Microsoft stated, “We have big ambitions for where we go next. Not only will we pursue further advances here, but we believe that orchestrating a range of specialized models serving different user intents and use cases will unlock immense value.”
MAI-Voice-1 speech model
It is able to generate an audio clip in just under one second on a single GPU.
Notably, Microsoft has been utilising its model to power Copilot Daily, providing an audio briefing regarding daily tips, news, and more, to create an audio interaction between two AI participants.
MAI-1-preview
The MAI-1-preview is the company’s first model trained completely in-house.
Additionally, it is particularly designed to offer a range of robust capabilities to consumers looking to benefit from models.
Availability of MAI-Voice-1 speech model, MAI-1-preview
Users can test the MAI-Voice-1 speech model in Copilot Labs. Meanwhile, MAI-1-preview is accessible for tests on LMArena, with expansion plans for certain text-based use cases on Copilot in the near future.
