Microsoft has introduced three new foundational AI models capable of generating text, voice, and images, marking a step in its effort to expand its in-house multimodal AI capabilities while maintaining its partnership with OpenAI. The models, developed by the company’s Microsoft AI research division, aim to compete with offerings from other AI labs and position Microsoft with a broader, self-built model stack.
Three Models Target Text, Voice, And Visual Generation
The release includes MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2, each focused on a different modality. MAI-Transcribe-1 converts speech into text across 25 languages and runs 2.5 times faster than Microsoft’s Azure Fast transcription service, according to the company. MAI-Voice-1 generates audio, producing up to 60 seconds of audio in one second and allowing users to create a custom voice. MAI-Image-2 is a video-generating model that expands Microsoft’s visual AI capabilities.
Models Rolled Out Across Foundry And Playground
MAI-Image-2 first appeared on March 19 through MAI Playground, a testing platform for large language models. Microsoft has now released all three models on Microsoft Foundry, while the transcription and voice models are also available through MAI Playground.
Superintelligence Team Leads Development
The models were developed by Microsoft’s MAI Superintelligence team, formed in November 2025 and led by Mustafa Suleyman. In a company blog post, Suleyman described the approach as “Humanist AI,” focusing on practical use and communication patterns. He added that more models are expected to be released through Foundry and integrated into Microsoft products and services.
Pricing Positioned Against Competing Models
Microsoft said pricing is a key factor in the release as competition in the large language model market increases. MAI-Transcribe-1 starts at $0.36 per hour. MAI-Voice-1 starts at $22 per one million characters. MAI-Image-2 starts at $5 per one million tokens for text input and $33 per one million tokens for image output. The company stated these prices are lower than comparable models from Google and OpenAI.
OpenAI Partnership Continues Alongside Internal Development
Despite expanding its own model lineup, Microsoft continues its partnership with OpenAI. Suleyman said in an interview with VentureBeat that the collaboration remains in place. He also told The Verge that a recent renegotiation allows Microsoft to pursue its own superintelligence research more independently.
Microsoft has invested more than $13 billion in OpenAI and integrates its models across various products through a multi-year partnership. The company follows a similar approach in hardware, developing its own chips while also sourcing from external suppliers.
Featured image credits: Flickr
For more stories like it, click the +Follow button at the top of this page to follow us.