Wan 2.7 Image is Alibaba's unified model for image generation and editing. Write a prompt to create a new picture, or upload up to 9 images and describe the change in plain words. It renders long text cleanly, shapes faces in fine detail and outputs 1K or 2K, with 4K text-to-image on Pro. No design software needed.
One model for creating and editing. Clear text inside the picture. Faces that look like real people. Turn a prompt or a stack of references into a finished visual without switching tools, then keep refining it in the same place until it is ready to publish.
Wan 2.7 Image handles text-to-image and image editing in the same flow. Create a product shot from a prompt, then upload it and ask for a new color, a new background or a cleaner layout. You never move between separate tools, so the look stays consistent from first draft to final asset. Ad concepts, landing page graphics, promotional banners and social posts move from idea to usable image in minutes.
Say goodbye to the generic AI face. Describe bone structure, eye shape, facial contour, makeup, hairstyle and accessories, and the portrait follows every detail. Each face looks intentional and distinct, with its own age, identity and styling, not a copy of a template. It suits profile pictures, avatars, beauty visuals and character art.
The model writes long passages inside images across 12 languages, including Chinese and English. Posters, menus, charts, formulas, tables, reports and infographics come out readable and stable. Put the exact words in quotes and the layout keeps them where you asked. Text-heavy designs need far less manual cleanup before they go live.
Upload up to 9 images and describe how to combine them. Put a person and a pet in one portrait, move a product into a new scene or borrow the style of one picture for another. Faces, objects and materials stay recognizable while lighting and perspective blend naturally. Composites that used to take hours of masking become a single prompt.
Change only what you name. Ask for a new word on a sign, a different date, a recolored glass or one less stray object, and the rest of the picture stays put. Fonts, shadows, reflections and textures match the original scene. It is ideal for retouching product photos and updating ready-made designs for a new season or market.
Re-shoot an existing picture from a new angle or through a new lens, such as a fisheye or a wider view. For new images, pick from eight aspect ratios, from square and portrait to cinematic 21:9 and ultra-wide 8:1 banners. The Standard version outputs 1K or 2K. Pro adds 4K for text-to-image work that needs print-level detail.
Create stunning AI images in three simple steps:
Open the workspace with Wan 2.7 Image already selected. For a new picture, describe the subject, scene, light and any exact text. To edit, switch to image to image and upload up to 9 JPG, PNG or WebP files, each up to 10 MB and at least 240 px per side.
Choose Standard for fast everyday work or Pro for higher detail. Set 1K or 2K, or 4K on Pro for text-to-image. New images can use eight aspect ratios, from 1:1 to ultra-wide 8:1; edits keep the shape of your uploaded image.
Confirm the credits and generate. Your image appears in your library, ready to preview and download in full resolution. Not quite right? Upload the result, describe one more change and run it again.
Both versions share the same prompts, references and editing skills, so you can switch at any time without rewriting anything. Standard is fast and light on credits, which makes it the right choice for drafts and daily content. Pro adds 4K text-to-image and extra polish for final assets.
| Feature | Standard | Pro |
|---|---|---|
| Best for | Fast everyday visuals, drafts and quick edits | Final campaign art, brand visuals and print work |
| Text to image | Yes, from one prompt | Yes, from one prompt |
| Image editing | Natural-language edits on uploaded images | Natural-language edits on uploaded images |
| Reference images | Up to 9 | Up to 9 |
| Resolution | 1K or 2K | 1K, 2K, or 4K for text-to-image |
| Aspect ratios | 1:1, 16:9, 4:3, 21:9, 3:4, 9:16, 8:1, 1:8 | 1:1, 16:9, 4:3, 21:9, 3:4, 9:16, 8:1, 1:8 |
| Text rendering | Long text in 12 languages | Long text in 12 languages, finer detail |
| Portraits | Fine control of face, makeup and hair | The same control with richer skin and fabric texture |
| Prompt length | Up to 5,000 characters | Up to 5,000 characters |
| Speed | Faster | Slower, more detailed |
| Images per run | 1 | 1 |
| Credits per image | 3 | 8 |
Answers to the questions people ask most before they create: editing, references, resolution, on-image text, prompts and credits. Short answers first, details after.
Wan 2.7 Image is Alibaba's unified AI model for image generation and editing. It creates pictures from text, edits uploaded images with plain instructions and combines up to 9 references in one result.
Yes. New accounts get free credits, and a Standard image costs only 3 credits, so you can test several prompts and edits before you choose a plan.
Pro gives finer detail and adds 4K for text-to-image, while Standard is faster and costs fewer credits. Both edit images, accept up to 9 references and render text in 12 languages, so most people draft on Standard and finish on Pro.
Yes. Upload a photo and describe the change, such as a new background, a different color or a replaced word. The model changes that part and keeps the rest of the picture intact, including faces, light and texture.
You can upload up to 9 images per edit. Use JPG, PNG or WebP files up to 10 MB each, with at least 240 px on every side. If a file breaks a rule, the workspace tells you exactly which one and why.
Standard outputs 1K or 2K, and Pro adds 4K for text-to-image. New images can use 1:1, 16:9, 4:3, 21:9, 3:4, 9:16, 8:1 or 1:8, which covers feeds, stories, banners and wide hero images. Edited images follow the uploaded picture.
Yes. It renders long, readable text in 12 languages, including Chinese and English, which suits posters, menus, charts and infographics. Put the exact words in quotes and keep each text block short for the best result.
Yes. Prompts can be in Chinese or English and can run up to 5,000 characters, so you can describe complex scenes, layouts and long on-image text in full detail.
Each Standard image costs 3 credits at 1K or 2K. Each Pro image costs 8 credits at 1K, 2K or 4K. Every run returns one image, and the workspace shows the exact cost before you generate.
Describe the face in concrete terms: bone structure, eye shape, contour, makeup, hairstyle and accessories. The more specific the description, the more distinct and natural the portrait looks. Add lighting and lens details for a photographic feel.
It is best for ad creatives, product photos, social posts, posters with text, portraits and avatars, and quick fixes to existing designs. Because creating and editing live in one tool, a small team can produce a full campaign without a long design cycle.
Write a prompt or upload your references. Get clean text, real-looking faces and sharp 2K or 4K detail in a single run, then refine the result until it is ready to share on your site, store or feed.
Create now