Microsoft AI, the tech giant's research division, announced the release of three new foundational AI models on April 2, 2026, as it accelerates its push to build an independent AI stack — even while maintaining its long-standing partnership with OpenAI.
The Three New Models
The three models were developed by Microsoft's MAI Superintelligence team, led by CEO of Microsoft AI Mustafa Suleyman, and are now available on Microsoft Foundry:
- MAI-Transcribe-1 — A speech-to-text model supporting 25 languages, 2.5× faster than Microsoft's Azure Fast. Starts at $0.36/hour.
- MAI-Voice-1 — An audio generation model. Generates 60 seconds of audio in one second. Supports custom voice creation. Starts at $22 per 1M characters.
- MAI-Image-2 — Image and video generation model, previously soft-launched on MAI Playground March 19. Starts at $5 per 1M tokens (text) / $33 per 1M tokens (image).
Why This Matters for Indian Businesses
- Lower cost AI models — Cheaper alternatives to Google and OpenAI.
- Multimodal capabilities — Opens new use cases for WhatsApp Commerce, support bots, and multilingual marketing.
- Azure integration — Plug directly into existing Azure workflows.
Microsoft's AI Independence Strategy
Despite investing over $13 billion into OpenAI, Microsoft is building its own AI stack in parallel. Suleyman: "At Microsoft AI, we're building Humanist AI — putting humans at the center, optimizing for how people actually communicate."
What's Next
Suleyman: "You'll see more models from us soon in Foundry and directly in Microsoft products and experiences."
Source: TechCrunch — April 2, 2026