Microsoft Launches MAI-1-Preview, MAI-Voice-1 AI Models: Features, Performance And Other Key Details

Microsoft is launching MAI-Voice-1, a new speech model that produces clear, expressive audio.

Microsoft (Photo by Sam Torres on Unsplash)

Microsoft has introduced two new in-house AI models, MAI-Voice-1 and MAI-1-preview. This marks a shift towards building its own artificial intelligence capabilities, reducing reliance on OpenAI.

MAI-Voice-1 is designed to generate audio quickly and efficiently. The MAI-1-preview model is now in public testing and ranks 13th on LMArena, a benchmark for language models.

In a blog post on Thursday, the company said its goal is to create technology that empowers people everywhere.

“At Microsoft AI (MAI), we believe AI should be used to empower every person on the planet. We are creating AI for everyone, a supportive, helpful presence always in the service of humanity,” the company said. Microsoft's AI division is known as MAI. 

“Since last year, we’ve been focused on building the foundation for this vision, with a world-class team and infrastructure. To fully meet our goals, MAI requires purpose-built models. Today, we’re excited to preview the first steps to making this a reality,” the company mentioned.

Also Read: Why AI Keeps Security Experts Awake At Night?

First, Microsoft is launching MAI-Voice-1, a new speech model that produces clear, expressive audio. It’s now available in Copilot Daily, Podcasts, and a new Copilot Labs feature. With support for both solo and group voices, it offers natural, high-quality sound. Microsoft sees voice as the next big step in making digital tools more personal and accessible.

“Voice is the interface of the future for AI companions and MAI-Voice-1 delivers high-fidelity, expressive audio across both single and multi-speaker scenarios,” it said.

MAI-Voice-1 delivers ultra-fast speech generation, producing a minute of audio in under a second on a single GPU. Users can explore expressive storytelling, create guided meditations, or build dynamic voice experiences on this model with just a simple prompt.

The other AI model released by MAI is known as MAI-1-preview. Microsoft labelled the model as an “in-house mixture-of-experts”, pre-trained and post-trained on 15,000 NVIDIA H100 GPUs.

MAI-1-preview is built to deliver clear, helpful responses by following everyday instructions with ease. It is being introduced gradually in select Copilot text features, where it will assist users across common tasks.

“We will continue to use the very best models from our team, our partners, and the latest innovations from the open-source community to power our products,” Microsoft said.

Watch LIVE TV, Get Stock Market Updates, Top Business, IPO and Latest News on NDTV Profit. Feel free to Add NDTV Profit as trusted source on Google.
GET REGULAR UPDATES
Add us to your Preferences
Set as your preferred source on Google