Microsoft keeps filling out its in-house model stack, and image generation is the latest piece. MAI-Image-2.5, announced at Build 2026, is the company’s strongest image model yet — it launched at number 2 for image editing on the Arena leaderboard, ahead of Nano Banana 2.
## What MAI-Image-2.5 does
The headline capability is identity-preserving image-to-image editing. The model keeps recognizable faces — along with hair, clothing, and full-body identity — consistent across changes in style, pose, and layout, which is aimed squarely at branded characters, spokespeople, and social campaigns. It also handles full-frame restyling like anime, color grading, film grain, and de-aging, and can restructure a shot by adding, removing, or repositioning objects and adjusting how people pose and interact. Microsoft reports a +75 point gain over MAI-Image-2 overall, with the biggest jumps in text rendering (+107) and cartoon/anime.
## Where to use it
The model ships in two flavors: MAI-Image-2.5 for maximum fidelity and MAI-Image-2.5-Flash for fast, scalable production. It is already live in PowerPoint, rolling out to OneDrive, and available on Microsoft Foundry for developers, pitched on quality per dollar.
Top AI Product 
Leave a comment