Wan2.7 Image (including enhanced version Wan2.7 Image Pro) is a professional AI image generation and editing unified model launched by Alibaba Tongyi Laboratory in 2026. It is positioned as a "controllable design level" cultural map tool. The core upgrade lies in the potential space mapping of text semantics and visual semantics. It focuses on three breakthroughs: portrait authenticity, color accuracy, and text rendering specialization. Blind evaluation of human preferences ranks first in China. This model completely bids farewell to the "AI standard face" and can generate "lifelike" portraits with micro details such as pores and fuzz. It supports "color palette" level color control such as Hex color value and reference image color selection, and achieves breakthrough 3K Token ultra long text rendering and 12 language mixing. It can stably output clear complex layout content such as Chinese characters, formulas, tables, etc. The Pro version further enhances the professional photography quality, optimizing for commercial portraits, movie level photographs, advertising posters, and other scenes. It can generate 8K ultra high definition images, accurately restore textures of materials such as velvet, metal, and skin, and support local editing, style transfer, and batch creation. It is widely used in e-commerce product display, film and television short drama storyboarding, self media content production, education and research image matching, and other fields, providing creators with an efficient solution from creativity to implementation, greatly reducing the threshold and cost of visual content production.
