project
Wan2.7-Image - An AI image generation and editing model launched by Alibaba Tongyi
Wan2.7-Image is an AI image generation and editing model developed by Tongyi Labs. It supports deep customization of character images (face shape, eye shape, bone structure, etc.), bidding farewell to the monotonous "AI standard face," and can accurately generate 4000+ characters and...
What is Wan2.7-Image?
Wan2.7-Image is an AI image generation and editing model developed by Tongyi Labs. It supports deep customization of character images (face shape, eye shape, bone structure, etc.), bidding farewell to the cookie-cutter "AI standard face." It can accurately generate over 4000 characters and multiple languages including simplified and traditional Chinese, English, Japanese, Korean, German, and French, eliminating the hassle of garbled characters. The model can precisely control brand colors through Hex color values, ensuring compliance with VI specifications. Wan2.7-Image is suitable for short drama creation, brand design, and other scenarios, and is now available on platforms such as Tongyi Wanxiang.
Main functions of Wan2.7-Image
-
Character customizationWan2.7-Image supports deep customization of details such as face shape, eye shape, bone structure and skin texture, and can generate personalized virtual images with high recognizability and natural texture.
-
Text generationThe model supports accurate rendering of ultra-long texts up to 4000 characters long, covering multiple languages including simplified and traditional Chinese, English, Japanese, Korean, German, and French. It can stably output tables, mathematical formulas, and multilingual mixed content.
-
Color controlWan2.7-Image innovatively introduces the "Color Control Palette" function, which supports direct input of Hex color values or uploading reference images to analyze color palettes. It can accurately set the brand's main color, auxiliary color and their proportions to ensure that the generated materials strictly comply with VI specifications.
-
Multi-image reference generation:It supports up to 9 images as references, maintaining strong consistency among multiple subjects.
- Interactive editingIt supports selecting a local area for precise modification, achieving pixel-level alignment between the intent and the AI.
How to use Wan2.7-Image
-
Regular usersYou can directly visit the Tongyi Wanxiang official website and generate an image by entering a description through the web interface.
-
DevelopersBy accessing the API through the Alibaba Cloud Bailian platform, Wan2.7-Image can be integrated into your application.
Key information and usage requirements of Wan2.7-Image
- Product PositioningTongyi Labs' AI image generation model boasts "more realistic human figures, more stable text, and more accurate colors."
- Supported languagesSupports multiple languages including Simplified and Traditional Chinese, English, Japanese, Korean, German, French, Spanish, and Italian.
- Input StandardsSupports natural language descriptions; for complex requirements, it is recommended to specify parameters such as facial features, color Hex values, and text content in detail.
The core advantages of Wan2.7-Image
- Breakthrough in character realismRejecting the cookie-cutter "AI standard face," it supports deep customization of face shape, eye shape, bone structure, and skin texture to generate virtual images with unique recognizability and natural texture, meeting the high requirements for character consistency in short dramas, brand IPs, and other applications.
- Text rendering accuracyIt boasts industry-leading ability to generate over 4000 long characters, supports mixed text output of multiple languages including Chinese, English, Japanese, Korean, German, and French, and can stably output tables, mathematical formulas, and dense text layouts, completely solving the pain point of "Martian text" garbled characters in AI-generated images.
- Color control precisionThe innovative "Color Control Palette" function supports direct input of Hex color values (such as #2C3E50) to define brand VI specifications. It allows setting the ratio of primary and secondary colors to ensure that the generated materials are completely consistent with the brand colors, eliminating color differences.
- Multi-image reference consistencyIt supports up to 9 images as style/character references, achieving strong consistency across multiple subjects and scenes, making it suitable for serialized content creation.
Comparison of Wan2.7-Image with similar competing products
| Comparison Dimensions | Wan2.7-Image | Midjourney | JiMeng AI |
|---|---|---|---|
| Text rendering | Supports 4000+ characters, mixed text in 13 languages, and stable output of formulas/tables. | The text often displays garbled characters or is corrupted, requiring post-processing. | Supports Chinese characters, but stability is limited for very long texts. |
| Color control | Supports precise Hex color value input and allows for the definition of brand VI specifications. | Relying on natural language descriptions, color accuracy is somewhat subjective. | It supports color picking from reference images, but lacks quantized Hex input. |
| Character consistency | You can specify face shape/eye shape/bone structure; 9 reference images maintain consistency across multiple subjects. | Multiple card draws are required; consistency depends on the Seed value or an external plugin. | Character references are supported, but the depth of customization for facial features is insufficient. |
| Interactive editing | Supports precise local editing with pixel-level alignment. | Partial editing is not supported; the entire image must be regenerated. | Supports smart canvas and partial redraw |
| Core advantages | Deep integration of accurate text and images, brand color accuracy, and consistent persona. | Top-notch artistic aesthetics and lighting quality, diverse styles | Strong Chinese semantic understanding and outstanding video generation capabilities |
| Applicable Scenarios | Brand materials, educational publishing, AI short dramas, e-commerce design | Artistic creation, concept design, illustration | Short videos, social media content, rapid creative ideas |
Application scenarios of Wan2.7-Image
-
AI short dramas and virtual idol creationWan2.7-Image can be used to create AI short dramas and virtual idols. By deeply customizing facial features and using multiple image references to maintain character consistency, it generates a unique virtual actor image that is highly recognizable and remains consistent across multiple episodes.
-
Brand VI and Marketing Material DesignThe model is suitable for brand visual identity system design and supports direct input of Hex color values to accurately lock the brand's standard color, ensuring that the color scheme of marketing materials such as posters and packaging is completely consistent with the company's VI specifications, thus completely eliminating color difference issues.
-
Educational Publishing and Knowledge VisualizationIn the field of education, it can stably render textbook illustrations or academic posters containing 4,000 characters, mathematical formulas, and mixed text in 13 languages, achieving a clear and accurate mixed text and image effect like printed materials.
-
Film storyboards and advertising storyboardsFor pre-production of film and television, the image generation function can produce logically coherent and stylistically consistent continuous storyboards, helping directors quickly visualize multi-angle shots of the same scene or the continuous action sequence of characters.