Key takeaways
- Microsoft launches MAI-Voice-1 and MAI-1-preview, advancing in-house AI capabilities.
- MAI-Voice-1 enables fast, expressive audio generation for interactive applications.
- MAI-1-preview is currently in public testing, leveraging 15,000 NVIDIA GPUs.
- Trusted testers can apply for API access to provide feedback for refinement.
- Microsoft aims to develop specialized AI systems for diverse user needs.
Microsoft AI has released its first in-house speech generation model, MAI-Voice-1, and begun public testing of its large language model MAI-1-preview, marking a step in the company’s push to develop purpose-built AI systems.
MAI-Voice-1 is already available in Copilot Daily and Podcasts, and is now accessible through Copilot Labs. According to Microsoft, the model can generate “a full minute of audio in under a second on a single GPU.” Microsoft says it is designed to enable expressive, multi-speaker audio for interactive use cases such as storytelling and guided meditations.
Alongside its speech model, MAI has started testing MAI-1-preview, an in-house mixture-of-experts model trained on about 15,000 NVIDIA H100 GPUs. The model is currently available for evaluation on LMArena, a platform for community testing, and will roll out gradually across Copilot text use cases.
“We have big ambitions for where we go next – model advancements, an exciting roadmap of compute, and the chance to reach billions of people through Microsoft’s products,” said Microsoft AI CEO Mustafa Suleyman.
“Voice is the interface of the future for AI companions,” the company said in its announcement. “We will continue to use the very best models from our team, our partners, and the open-source community to power our products.”
Microsoft confirmed that trusted testers can apply for API access to the model. The company said feedback from early users will help refine the system before wider deployment.
The new models are part of MAI’s long-term goal of orchestrating specialised AI systems tailored to different user intents. “We’re focused on creating applied AI as a platform for trusted products that understand each of our unique needs,” the team said.
The announcement also highlighted infrastructure expansion, noting that MAI’s next-generation GB200 cluster is now operational. Microsoft said this will support future model development at scale. MAI described itself as a “lean, fast-moving lab” and is seeking new hires.
