Microsoft AI, the technology giant’s research division, disclosed on Thursday the debut of three groundbreaking AI models capable of generating text, audio, and visuals.
This launch highlights Microsoft’s ongoing dedication to developing its proprietary suite of multimodal AI models, positioning itself as a strong competitor to other AI research entities, despite its continued collaboration with OpenAI.
MAI-Transcribe-1 offers speech-to-text transcription across 25 languages, performing 2.5 times faster than Microsoft’s Azure Fast service, according to the company’s press release. MAI-Voice-1 is an audio creation model, enabling users to generate 60 seconds of audio in just one second and customize voices. MAI-Image-2, meanwhile, functions as a video generation model.
MAI-Image-2 initially launched on MAI Playground, a new large language model testing platform, on March 19. Now, all three models are accessible on Microsoft Foundry, with the transcription and voice models also available on MAI Playground.
These models were developed by Microsoft’s MAI Superintelligence team, an AI research group led by Mustafa Suleyman, CEO of Microsoft AI, which was established and announced in November 2025.
“At Microsoft AI, our mission is to build Humanist AI. We approach AI model creation with a distinct perspective – centering humans, optimizing for natural communication, and training for practical applications,” Suleyman articulated in the blog post. “Expect to see more models from us soon, available both in Foundry and directly within Microsoft products and experiences.”
In an increasingly competitive LLM market, MAI aims for these models to stand out due to their lower cost compared to offerings from Google and OpenAI, as stated in the company’s blog post.
MAI-Transcribe-1 begins at $0.36 per hour. MAI-Voice-1 is priced from $22 per 1 million characters, while MAI-Image-2 starts at $5 for 1 million tokens for text input and $33 for 1 million tokens for image output.
Despite introducing its own models, Suleyman reiterated Microsoft’s commitment to its alliance with OpenAI in an interview with VentureBeat – although a recent adjustment to that partnership enabled Microsoft to actively pursue this superintelligence research, Suleyman informed The Verge.
Microsoft has injected over $13 billion into the AI research lab, hosting its models across various products through a multi-year agreement. Microsoft adopts a similar approach with chips; it both manufactures its own and sources them from external providers.