Alibaba Cloud has updated and provided the ability to call Wan2.6-Image (wan2.6-image) in Model Studio, positioning it as a "production-ready" image generation and editing model. According to official information, the model supports staggered output of images and texts, and can generate content containing text and multiple images in a single request, which can be used for narrative visual expressions such as step-by-step tutorials and storyboards.
At the generation and editing level, WAN 2.6-image supports 1–3 reference images to participate in the generation, which can realize the creation of multi-image conditions such as style reference, background replacement, and element fusion, and provides image editing capabilities to reduce repeated redrawing. The interface side uses asynchronous task mode to return the results, the output format is PNG, and the "AI Generated" watermark can be turned on on demand. Text prompts are available in both Chinese and English and have length limits. Commercial use still needs to pay attention to the compliance of material licensing, portraits and brand elements, and avoid using unauthorized reference images directly for commercial use.
FAQs
Q: What is the difference between Wan2.6-Image and ordinary Wensheng Diagram?
A: In addition to Wensheng Diagram, it also supports image editing, staggered output of graphics and texts, and fusion and replacement with multiple reference images.
Q: Can Wan2.6-Image combine the style and background of multiple reference images?
A: Yes, the interface example supports passing in multiple images at the same time, allowing the model to borrow elements such as style and background according to the instructions.
Q: What content is the image and text output of Wan2.6-Image suitable for?
A: It is suitable for "step-by-step tutorials", "story storyboards", "poster drafts with instructions", etc., which require text and pictures to be interspersed.
Q: Does Wan2.6-Image support commercial consistency requirements?
A: The official publicity emphasizes a more stable style and subject consistency for commercial scenarios, but the specific effect still needs to be tested and verified in its own materials and processes.
