MAI-Image-2.6

MAI-Image-2.6 is an image generation and editing model from Microsoft AI that works from text, images, web context, and visual references. Multi-reference editing lets it combine people, products, styles, and scenes drawn from separate inputs, while web grounding pulls current information off the web so generated visuals reflect real, present-day context rather than training-set memory. The model supports dynamic aspect ratios and resolutions up to 1.5K, and ranks No. 2 for text-to-image and No. 1 for image editing on Artificial Analysis, with No. 2 Arena placements in both categories.

Production image workloads are gated less by peak quality than by what that quality costs at volume — the strongest models have been too slow and too expensive to sit inside a high-throughput pipeline. A companion variant, MAI-Image-2.6-Flash, targets exactly that constraint: it generates images 2.8× faster than GPT-Image-2-Medium with 72% greater efficiency, at less than half the price of the flagship. Both variants are available in public preview through Microsoft Foundry and the MAI Playground.