Microsoft Challenges OpenAI with 3 New "In-House" AI Models

Microsoft Takes on AI Rivals With Three New Foundational Models:

The AI Wars Just Changed: Microsoft Drops 3 Powerful New Foundational Models

In a bold move to assert its dominance in the rapidly evolving AI landscape, Microsoft has officially unveiled three new foundational AI models — marking a decisive step in the company's mission to build its own full-stack multimodal AI platform. Released by Microsoft AI, the company's dedicated research lab, these models span speech transcription, voice synthesis, and image generation, putting Microsoft squarely in competition with industry giants like Google, OpenAI, and Anthropic.

This announcement signals more than just new product launches. It signals Microsoft's deepening commitment to building proprietary AI infrastructure — even as it maintains its multi-billion-dollar partnership with OpenAI.

What Are the Three New Microsoft AI Models:

• MAI-Transcribe-1 — Speech-to-Text Redefined. The first of the trio is MAI-Transcribe-1, a state-of-the-art speech transcription model capable of converting spoken language into text across 25 different languages. What makes it exceptional is its speed — it is 2.5 times faster than Microsoft's existing Azure Fast transcription offering. Available at $0.36 per hour, it's priced competitively to challenge dominant players in the AI speech-to-text market.

• MAI-Voice-1 — Custom Voice at Unprecedented Speed. Next is MAI-Voice-1, a powerful audio-generating model that can produce 60 seconds of audio in just one second — a feat that sets a new benchmark for real-time voice synthesis. Beyond speed, the model empowers users to create fully customized AI voices, unlocking new possibilities for content creators, developers, and enterprises. Pricing starts at $22 per 1 million characters.

• MAI-Image-2 — Vision AI Meets the Foundry. Rounding out the trio is MAI-Image-2, a video and image generating model that first debuted on MAI Playground on March 19, 2025 — Microsoft's newly launched large language model testing environment. Now available on Microsoft Foundry, it starts at $5 per 1 million tokens for text input and $33 per 1 million tokens for image output.

The Team Behind the Models:

Microsoft's MAI Superintelligence team is the powerhouse driving these innovations. Announced in November 2025 and led by Mustafa Suleyman, CEO of Microsoft AI, the team is dedicated to building AI that is both cutting-edge and fundamentally human-centric.

In a company blog post, Suleyman articulated a clear philosophical direction: "At Microsoft AI, we're building Humanist AI. We have a distinct view when creating our AI models — putting humans at the center, optimizing for how people actually communicate, training for practical use." He also confirmed that more models are coming soon — both to Foundry and directly into Microsoft's consumer products.

Where Can You Access These Models:

All three models are now available on Microsoft Foundry, Microsoft's enterprise AI development platform designed for building, testing, and deploying AI-powered applications.

Additionally, MAI-Transcribe-1 and MAI-Voice-1 are available in MAI Playground — Microsoft's newly launched LLM testing environment that gives developers and creators early access to experiment with the latest models.

MAI-Image-2 was the trailblazer of this release, having first launched on MAI Playground back on March 19, giving early adopters a preview of Microsoft's visual AI capabilities before the full rollout.

Competing on Price: Microsoft's Smart Market Strategy:

Cost efficiency is one of Microsoft's most compelling differentiators in this crowded AI market. As the company noted in its blog post, the MAI models are priced below comparable offerings from Google and OpenAI — a strategic move to attract developers, startups, and enterprises who need powerful AI at a more accessible price point.

At a glance, here's the pricing breakdown:

• MAI-Transcribe-1: Starting at $0.36 per hour.

• MAI-Voice-1: Starting at $22 per 1 million characters.

• MAI-Image-2: Starting at $5 per 1M text input tokens / $33 per 1M image output tokens.

Microsoft and OpenAI: Partnership, Redefined:

Despite developing its own foundational models, Microsoft has not stepped away from OpenAI. In an interview with VentureBeat, Suleyman reaffirmed the company's commitment to the OpenAI partnership — while also acknowledging that a recent renegotiation of that partnership was what truly opened the door for Microsoft to pursue its own superintelligence research agenda.

Microsoft has invested more than $13 billion into OpenAI and continues to host its models across Microsoft products through a multi-year partnership. This dual strategy mirrors Microsoft's approach to hardware — the company both produces its own AI chips and sources them from external partners like NVIDIA, ensuring flexibility, resilience, and a competitive edge in the AI infrastructure race.

Why This Matters for the Future of AI:

The launch of MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 is not just about three new products. It's a statement of intent. Microsoft is building an end-to-end AI stack — from chips and infrastructure to models and applications — that positions it as an independent force in the global AI arms race.

For developers, businesses, and consumers, these models represent a major expansion of accessible, affordable, and high-performance AI tools. Whether you need fast multilingual transcription, lifelike custom voice generation, or scalable image synthesis, Microsoft's MAI suite is engineered to deliver — and to do so at a price that challenges the status quo.

Final Thoughts:

Microsoft is playing a long, calculated game in the AI space. With the MAI Superintelligence team now firmly established, a clear pricing strategy that undercuts rivals, and a platform ecosystem in Microsoft Foundry and MAI Playground, the company is building the rails for the next generation of AI applications.

As Mustafa Suleyman put it, "You'll see more models from us soon." In a market defined by speed, scale, and innovation, Microsoft is making sure it won't be left behind.