On September 24, Alibaba's AI flagship application, Quark, officially launched its new AI creation platform "Zao Dian." The platform integrates two core capabilities: AI image generation and AI video generation. It is the first platform in China to support synchronized audio-visual video generation by integrating Tongyi Wanxiang Wan 2.5. In terms of image generation, "Zao Dian" has introduced the globally leading image generation model Midjourney V7. Combined with Quark's self-developed technology, "Zao Dian" provides users with a one-stop creative experience that includes both images and videos.

image.png

The integration of Tongyi Wanxiang Wan 2.5 is a major highlight of the platform. Relying on the latest model capabilities, "Zao Dian" can generate voice, sound effects, and background music that match the visuals, supporting up to 10 seconds of video, allowing for more complete storylines in a single generation; it supports 1080P high-definition video at 24 frames per second, with image quality capable of meeting the requirements of cinematic scenes. The model has strong instruction-following capabilities, enabling complex continuous control such as camera movement during video generation.

In terms of AI image generation, "Zao Dian" fully supports multiple core capabilities of Midjourney V7, including multi-image reference, style codes, and detailed parameter control, which can meet the advanced needs of professional creators in composition, style expression, and detail refinement. Combined with Quark's self-developed capabilities, "Zao Dian" excels particularly in Asian portrait generation, Chinese text generation, and traditional Chinese elements, suitable for applications in e-commerce, graphic design, and film and game industries.

During the internal testing phase, "Zao Dian" attracted a large number of professional AIGC creators, producing popular works such as "Medieval Oil Painting Chorus," "Cultural Relics Eating Mooncakes," and "Giant Cat Traveling." The mouth movements, actions, expressions, and music in the videos are seamlessly integrated. One creator said, "Midjourney generates images, and then Wan 2.5 generates videos. The image quality is delicate and stable, especially this time the upgrade of the video model allows for synchronized output of audio and video, greatly improving the efficiency of creation."

image.png

Creative content created by creators on Quark's "Zao Dian"

Currently, users can directly access zaodian.quark.cn, or enter "Zao Dian" through the "AI Image Generation" feature on Quark's PC version. From now until September 30, all users can enjoy a 7-day free trial of the video generation capability of Tongyi Wanxiang Wan 2.5.

On the Quark APP, "Zao Dian" provides a lightweight creative experience, meeting the needs of more ordinary users. Users can not only achieve fun features like "one-sentence photo editing" and "AI creative videos," but also meet daily photo editing needs such as filters, beauty enhancement, and removing passersby; it also supports popular AI scenarios such as changing hairstyles, backgrounds, style conversion, and 3D figurine generation.