Tongyi Wanxiang
An AI creative creation platform under Alibaba's Tongyi, officially offering text-to-image, image-to-image, text-to-video, image-to-video, and image editing capabilities. Wanxiang 3.0 supports native 30-second duration generation with multi-material reference and is targeted at productivity scenarios.
Updated: 2026-10-26
What is Tongyi Wanxiang?
Tongyi Wanxiang (Wanxiang) is an AI creative creation platform under Alibaba's Tongyi brand. Its official website positions it as a "leading AI video and image generation model", aiming to lower the threshold for creation through artificial intelligence technology. It covers two major areas: images and videos. The creation scenarios listed on the official website include text-to-image, image-to-image, text-to-video, image-to-video, and image editing. The current flagship product is Wanxiang 3.0, which the official claims has capabilities such as native 30-second duration, generation with reference to any modality, and precise video editing. The platform also provides open-source models and API access.
Who is it for?
- Creative teams that produce commercials, concept demonstrations and short films and need to deliver multiple versions of proposals in one go
- E-commerce and marketing professionals who generate main video clips and promotional materials using reference images of products
- Content creators who need consistency in characters or scenes, using multi-material references and consistency preservation to reduce repeated trial-and-error
- Designers working on design proposals and storyboards who create dynamic previews before moving on to formal production
- Developers who want to integrate generation capabilities into their own systems can access the official API platform and open-source models on the official website
Key features
6Text-to-Image and Image-to-Image
The official website lists text-to-image and image-to-image as basic creation scenarios, allowing users to start from text descriptions or generate new visuals based on existing images.
Text-to-Video and Image-to-Video
Supports generating videos from text or images. Tongyi Wanxiang 3.0 is officially claimed to support a native 30-second duration, enabling more complete narrative capabilities and greater creative freedom, and also provides an intelligent duration function.
Arbitrary Modal Reference Generation
The official claims it supports up to 20 material references for generation, and also enables parsing of complex documents and web pages, allowing reference materials to be directly converted into visual content.
Pixel-level Consistency Maintenance
The official claims it can accurately restore and replicate reference details, enhancing delivery certainty for productivity scenarios.
Precise Video Editing
Video editing supports two modes: instruction-based editing and reference-based editing, which the official claims can improve creation efficiency.
Immersive Audio-Visual Experience
The official claims that the sense of realism, picture quality and sound effects have been comprehensively enhanced, creating a stronger audio-visual impact.
How to use Tongyi Wanxiang
- First, prepare all reference materials: Wanxiang 3.0 supports a maximum of 20 reference materials.
- Include several images for characters, costumes, scenes, and styles.
- Combined with pixel-level consistency maintenance, the subject consistency will be significantly better than pure text descriptions.
- Choose the duration as needed.
- the official states that it supports native 30 seconds and an intelligent duration function.
- Before officially creating a long version, it is recommended to first use a shorter content to verify the style and consistency to avoid consuming credits on trial and error.
- When modifications are needed, prioritize using video editing commands and reference editing, which is more cost-effective than regenerating.
- For batch or automated production, use the API platform.
- You can also follow the official website's model open-source and GitHub entrances.
- specific capabilities are subject to official releases.
- Generation is billed by credits or subscription.
- prices and free credits are subject to the description on the official website's "Product Pricing" page.
- When involving portrait and brand materials, be sure to confirm authorization before commercial release.
Traffic & popularity
Data month Sep 2026
Tongyi Wanxiang receives about 819.1K visits per month.
Monthly visits
12 monthsTraffic sources
- Direct66.0%
- Search18.4%
- Referrals12.7%
- Display ads1.5%
- Social1.5%
Top countries
- China89.9%
- Hong Kong, China3.0%
- United States2.2%
- Taiwan, China1.0%
- Singapore0.5%
Estimated from third-party traffic data.
In these collections
Similar tools
ViduThe AI video generation platform launched by Shengshu Technology features multi-subject reference video generation as its core strength. It excels particularly in character consistency and 2D animation rendering, and has been used by many AIGC directors in commercial projects such as manhua dramas and advertisements.
Miaohui AIMiaohui AI is a free AI painting generation platform under Shanghai Yuanxiang Information Technology. Officially, it supports text-to-image, image-to-image and AI video generation, covering styles such as character portraits, landscape photography, product display and artistic illustrations, and provides e-commerce image and prompt word assistance functions.
Fusion BrainFusion Brain is an all-in-one AI creation platform covering capabilities such as text-to-image generation, image editing and video generation. The official website is temporarily inaccessible, and the specific functions and pricing shall be subject to the official subsequent announcement.
Hailuo AIThe AI assistant and video generation platform under MiniMax provides text-to-video and image-to-video capabilities based on its self-developed video model. It stands out for its smooth motion and strong subject consistency, and also integrates general AI functions such as intelligent Q&A and voice.
HaimetaHaimeta is a creative platform focusing on "Human Vision Meets AI Art". According to the official statement, you can generate images, videos, 3D content and immersive worlds with just one sentence. It offers text-to-image, image-to-image, text-to-video, image-to-3D and a variety of popular special effects.
ToMovieeToMoviee is an AI content creation platform launched by Wondershare Technology. Officially, it claims that videos, images and music can be generated with one sentence. It covers modules such as text-to-video, image-to-video, video continuation, text-to-music and text-to-speech, and supports precise control of shots and emotions.