【#Tech24H】Alibaba has released Wan2.7-Image, a unified image generation and editing model, aiming to address current pain points in AI image generation such as fatigue over “standard faces” and “color shifting”artifacts. The model elivers diverse, lifelike facial generation while also enabling precise color control. To move beyond the monotonous “AI face”, Wan2.7-Image enhances the virtual avatar customization function, supporting full-scale personalization from bone structure, eye shape, to facial features. For example, users can change face shapes (oval, round, square, oblong, etc.) and eye characteristics (almond-shaped, deep-set, round, phoenix-eyed, etc.) via prompt adjustments, achieving “thousands of faces”. The model features full-chain capabilities including text-to-image, image-to-multiple-image, instruction-based image editing, and interactive editing.