Today, we are excited to announce three powerful new MAI models that set new standards for performance and value.
All three models MAI-transcribe-1, MAI-voice-1, and MAI-image-2 are available now in Microsoft Foundry. Additionally, they can be accessed in the MAI Playground, where available.
We are introducing MAI-Transcribe-1 along with MAI-Voice-1 and MAI-Image-2. Each model delivers top-quality performance and fast delivery, now at highly competitive prices, building on our ongoing commitment to innovation.
To start exploring, you can access all three models now in Microsoft Foundry and MAI Playground.
First, MAI-Transcribe-1 provides advanced text speech-to-text transcription for the 25 most widely used languages, as measured by the FLEURS benchmark. Designed for high quality and real-world conditions, it offers batch transcription speeds 2.5 times faster than Microsoft Azure Fast. MAI-Transcribe-1 is both exact and fast, and it’s now available in Foundry at the best price-performance among major cloud providers.
Next, MAI-Voice-1 is our leading voice generation model. It creates natural, realistic speech with nuance, emotional range, and expression while keeping the speaker’s identity clear even in long-form content.
Today, we are adding the LT to safely and securely create your own custom voice in Microsoft Foundry with just a few seconds of audio. MAI-Voice-1 can change how easily developers can build high-quality, high-speed voice experiences and Voice agents.
Thanks to its capabilities, the model can generate 60 seconds of audio in just one second, and its efficient GPU use keeps costs low. Try it yourself with Co-Pilot audio expressions or Co-Pilot podcasts to experience the difference.
Moving on, MAI-image-2 delivers much faster image generation on Copilot, ranking among the top three model families on the Arena AI leaderboard. Users see at least twice the speed on Foundry and Copilot, while maintaining the same quality based on real-world data. It’s also being rolled out in Bing and PowerPoint.
Customers are already using MAI, Hyphen Image, and Hyphen 2 for creative projects. WPP, one of the world’s largest marketing and communications groups, is one of the first enterprise partners to build at scale with MAI-Image-2.
MAI-image-2 is a genuine game changer; it’s a platform that not only responds to the nuanced details of creative direction but also deeply respects the sheer craft involved in generating real-world, campaign-ready images, said Rob Reilly, Global Chief Creative Officer, WPP. WPP possesses some of the best creative talent in the world, and MAI-image-2 is making them even better.
MAI Models: Better, Faster, and Cheaper Than Our Competitors
We are rolling out these advanced models to support both our customers and business products. Our Microsoft Foundry customers can now benefit from improved quality, speed, and efficiency at competitive prices.
- MAI- transcribe-1 is available starting from $0.36 per hour
- MAI-voice-1 starts at $22 for every 1M characters.
- Mai-image-pricing begins at $5 per 1M tokens for text input and $30 per 1M tokens for image input.
Access these models now on Microsoft Foundry and MAI Playground start unlocking advanced AI capabilities today.
From today, all developers can use MAI models, including MAI-transcribe-1, through Microsoft Foundry. You can also try them out in the MAI Playground, currently only available in the US.
If you’re interested in MAI models but don’t have access to Foundry, take action now. Fill out this form today, and our team will reach out and help you get started.
Our models excel at every level.
At Microsoft AI, we focus on building humanist AI. Our approach puts people first, ensuring our models reflect reality, communicate effectively, and are trained for practical use. More models will be available soon in Foundry and in Microsoft products.
Consistent with our commitment to safe and responsible AI, these MAI models were developed, tested, and rigorously ready‑teamed. Through Microsoft Foundry, developers get built‑in guardrails, governance, and enterprise‑grade controls designed to support safe, compliant deployment at scale.
Source: Today we’re announcing 3 new world class MAI models, available in Foundry
