We have seen a bunch of amazing image models in the past few months. They are getting a lot better at generating and editing images. Take Wan2.7-Image for instance: it can generate realistic faces, giving you control over eyes, bone structure, and other aspects. This model offers text support across 12 languages.
1/8 Meet Wan2.7-Image — Our unified model for image generation and editing.
One model that generates, edits, and understands images:
Realistic faces with full control over bone structure, eyes, and contour
Color Palette with HEX codes and reference image extraction
3K-token text… pic.twitter.com/HlOixPSu5B— Wan (@Alibaba_Wan) April 1, 2026
This model can generate up to 12 consistent images in one generation. It lets you customize faces at a granular level. You can select the regions you want to edit. You can already access this model through the API or online interface.
[HT]
