Experiment

MAI-Image-2

MAI-Image-2 is a state-of-the-art text-to-image generation model from Microsoft AI (MAI), built to support high-quality creative and production workflows. Designed in collaboration with photographers, designers, and visual storytellers, the model prioritizes realism, reliable text rendering, and expressive range—enabling creators and enterprise developers to generate images that feel intentional, grounded, and ready for use.

Key capabilities include photorealistic image generation with natural lighting, accurate skin tones, and environments that feel genuinely lived in. It also offers reliable text rendering across a wide range of fonts, sizes, and orientations, making it well suited for posters, labels, signage, and design work. In addition, it supports expressive and imaginative creation, whether the goal is cinematic, surreal, or highly detailed visual concepts. Built with creators and developers in mind, it helps minimize post-processing and enables faster, more efficient iteration.

MAI-Image-2 is available for experimentation in early access and supports enterprise scenarios through private preview API access for select customers. The model is designed for integration into products and services at scale and will be available through Microsoft Foundry soon, with planned integration into Copilot and Bing Image Creator.