Copilot users get better photo lighting and in-image text for slides at no extra cost; developers pay per token through Foundry.

Kenning's take

What it does

MAI-Image-2 is a text-to-image model from Microsoft AI, built with input from photographers and designers. Microsoft says it improves natural lighting, skin tones and in-image text, which helps with posters, slides and diagrams. It runs in Copilot, with rollouts in Bing Image Creator and PowerPoint, and developers can use it in Microsoft Foundry and the MAI Playground; Microsoft has since released MAI-Image-2.6.

• Photorealistic lighting and skin tones
• Reliable text in images for posters and diagrams
• Detailed, surreal scene generation
• At least 2x faster generation on Foundry and Copilot (Microsoft figure)

Who it suits. Copilot users and developers who need realistic images or images with legible text.

Key facts

Category

AI models

Part of

Microsoft Copilot

Made by

Microsoft AI

Based in

United States

Released

19 March 2026

How to get it

in Copilot, rolling out in Bing Image Creator and PowerPoint, in the MAI Playground, and in Microsoft Foundry for developers

Works on

Web, iOS, Android, API

Free plan

Available in Copilot

API price

from $5 per million text input tokens and $33 per million image output tokens on Microsoft Foundry

FAQ

Facts from the maker's own pages and the sources listed. Checked 2026-10-11.