Microsoft takes on Google and OpenAI with its own AI models

From transcribing boardroom chatter to cloning voices in seconds, Microsoft's MAI model trio is here, and it is priced to make rivals sweat.

Microsoft takes on Google and OpenAI with its own AI models

Microsoft just shipped its own AI models, and they’re coming for OpenAI and Google. The company has publicly released three proprietary models: MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2. The models are available via the Microsoft Foundry platform and the MAI Playground. 

We’re bringing our growing MAI model family to every developer in Foundry, including …

· MAI-Transcribe-1, most accurate transcription model in world across 25 languages
· MAI-Voice-1, natural, expressive speech generation
· MAI-Image-2, our most capable image model yet

Start… pic.twitter.com/p0DZZcAUZ4

— Satya Nadella (@satyanadella) April 2, 2026

So, what can Microsoft’s AI models actually do?

The trio covers a variety of use cases: listening, speaking, and seeing. MAI-Transcribe-1, for instance, handles speech-to-text across 25 languages and is 2.5 times faster than Microsoft’s own Azure Fast offering. It’s worth mentioning that the audio model was built by a team of 10 people. 

MAI-Voice-1 can produce 60 seconds of natural-sounding audio in just a second. It also supports creating custom voices from just a short audio clip. MAI-Image-2, meanwhile, has already secured a position in the top three on the Arena.ai image generation leaderboard. Rollouts are currently underway in Bing and PowerPoint. 

None of this has happened overnight, though. Until October 2025, the company was contractually restricted from building its own frontier AI by none other than OpenAI. Both companies signed a deal in 2019 that gave Microsoft a license to OpenAI’s models in exchange for facilitating OpenAI’s cloud infrastructure. 

Is Microsoft ready to cut ties with OpenAI?

However, the deal also barred Microsoft from making its own AI models. Once that changed, Microsoft released its own AI models, the ones that quietly powered Copilot and Teams behind the scenes. The models are available for any developer on Foundry to build with. 

Not yet. Mustafa Suleyman, the CEO of Microsoft AI, has reaffirmed the company’s commitment to its OpenAI partnership, even as those models signal a parallel strategy taking shape. The pricing is quite sharp as well. All three models are priced below the comparable offerings from Amazon and Google

If these models perform well, the MAI family could quietly become the backbone of Microsoft’s entire AI product portfolio.