← Back to Blog

Microsoft takes on AI rivals with three new foundational models

B
Baseinsider Team
Author
Articles
Apr 04, 2026
2 min read
60 views
Microsoft takes on AI rivals with three new foundational models
Microsoft AI has launched three new foundational models that generate text, voice, and images, marking a strategic push to compete in the multimodal AI space. The models include MAI-Transcribe-1 for fast transcription, MAI-Voice-1 for customizable audio, and MAI-Image-2 for video and image generation. These releases highlight Microsoft's drive for accessible, human-centered AI solutions alongside ongoing partnerships with OpenAI.

Microsoft takes on AI rivals with three new foundational models

Microsoft AI Unveils Three Next-Gen Multimodal Models

Microsoft AI has rolled out three advanced foundational models designed to generate text, voice, and images, further intensifying competition in the artificial intelligence arena. This move underscores Microsoft's commitment to building a comprehensive suite of multimodal AI tools, even as it continues to partner with OpenAI.

An Overview of the New Models

  • MAI-Transcribe-1: This speech transcription model supports 25 languages and delivers results 2.5 times faster than Microsoft's previous Azure Fast offering.
  • MAI-Voice-1: Capable of generating 60 seconds of audio in just one second, this tool lets users create personalized voice outputs efficiently.
  • MAI-Image-2: Originally introduced on MAI Playground, this model enables video and image generation, expanding creative possibilities for AI-powered content.

Access and Pricing

All three models are now available on Microsoft Foundry, with the transcription and voice solutions also integrated into MAI Playground. MAI-Transcribe-1 starts at $0.36 per hour, MAI-Voice-1 is priced at $22 per million characters, and MAI-Image-2 costs $5 per million tokens for text and $33 per million tokens for image output. Microsoft’s pricing aims to be more competitive than offerings from Google and OpenAI.

Development and Vision

The models were created by the MAI Superintelligence team led by Mustafa Suleyman, Microsoft AI’s CEO. According to Suleyman, the company is focused on “Humanist AI”—placing people at the center of AI model design for practical, real-world applications. Microsoft expects to introduce further models within its Foundry platform and integrate them across its products in the near future.

Continued Partnership with OpenAI

Despite this independent innovation, Microsoft reaffirms its collaboration with OpenAI, supporting a diversified approach to AI advancement, including investing heavily in related research and also sourcing technologies from other leading sources.

To read the full original article, visit TechCrunch: Microsoft takes on AI rivals with three new foundational models.

Share this article

Twitter Facebook LinkedIn

Related Articles

Meta’s New Reality: Record High Profits. Record Low Morale

Meta’s New Reality: Record High Profits. Record Low Morale

Meta is experiencing record profits but faces extreme employee dissatisfaction following repeated layoffs, tighter performance tracking, and controversial AI-related initiatives. Workers report low morale, concerns over privacy, and a growing push for unionization, reflecting a company culture increasingly defined by fear and uncertainty, despite ongoing investment in AI development.

May 18, 2026 • 3 min read