Microsoft launches MAI Image 2 model for Copilot and Bing Image Creation

Microsoft has launched its second-generation AI image generation model, the MAI Image 2, which is said to generate more realistic images and generate text that’s easier to read. 

To achieve the level of realism, the company collaborated with photographers, designers, and visual storytellers during development to refine its performance.

The main goal for the MAI Image 2 model is to generate an image consisting of natural lighting, accurate skin tones, and realistic environments, which eventually helps minimise follow-up editing after the images are generated, saving users time. 

Microsoft’s blog post said, “MAI-Image-2 is built for creatives who want images that feel like they exist in the world, with natural light, accurate skin tones, environments that feel lived-in. Creatives can now spend less time fixing in post-production and more time making.”

Alongside photorealism, Microsoft has also refined in-image text generation, and it claims to generate accurate text for infographics, slides, diagrams, and more.

The MAI-Image-2 is available for preview on MAI Playground, and it's also rolling out to oll out on Copilot and Bing Image Creator. Microsoft aims to provide seamless workflows, enabling designers, marketers, and developers to generate high-quality visuals directly within applications they already use.

In addition, for API access, it is currently available for selected Microsoft customers. The company also plans wider access, which is expected to come soon through Microsoft Foundry for developers.

With the announcement, the tech giant also announced that its MAI-Image-2 has ranked #3 among AI labs for text-to-image generation globally, as per the Arena.ai leaderboard. 

While the results look pleasing, Microsoft is competing with some of the leading AI image models, including Google’s Nano Bana 2, OpenAI’s DALL-E 3, Midjourney 6.1, and others.