On February 19, 2024, Beijing Zhipu Huazhang Technology Co., Ltd. officially announced an important strategic cooperation. The company's multimodal creation platform "Qingying" has reached in-depth cooperation with the AI character creation platform "NianTa", and the two parties will jointly explore the complete industrial chain from AI character design to animation production. The focus of this cooperation is to promote the practical application of IP consistency technology, aiming to provide two-dimensional users with a better content creation experience. By integrating the technological advantages of both parties, cooperation will significantly improve the quality of generation effects and the speed of reasoning, bringing users an unprecedented creative experience.
As the core technical support of this cooperation, the "Qingying" platform has achieved breakthrough progress in multiple technical dimensions based on the latest upgraded video model CogVideoX-2 of Zhipu. The model has been comprehensively optimized in terms of architecture design, training methods and data engineering, which has improved the basic model capabilities of graph-generated videos by 38%. It is particularly worth mentioning that the CogVideoX-2 model can support the main body to perform large movements while maintaining picture stability, and is particularly outstanding in command compliance and artistic style control. On the first day of cooperation, the text and image generation video functions provided by "Qingying" for the "Nipai" platform have achieved remarkable results. The number of video generation in a single day exceeded 40,000 times, and the number of Tokens calls reached hundreds of millions.
As a top AI character creation platform with more than one million monthly active users, "Pin Ta" provides users with rich creative scenarios. Users can not only participate in scripts on the platform, experience script killing, but also create novels, comics, and even social interactions. Through this cooperation, users of "Pin Ta" can generate high-quality video content by simply using simple text or picture prompts, which greatly lowers the threshold for content creation and truly realizes the creative concept of "Pin Ta into a Film". This innovative functional integration provides content creators with more convenient and efficient creative tools.
This strategic cooperation marks the successful implementation of Zhipu multimodal model in commercial scenarios, and also shows Zhipu's rapid growth trend in the field of multimodal commercialization. Data shows that since the end of 2024, the average daily increase in the number of Zhipu multimodal series models has exceeded 1,000% compared with 2024, providing strong technical support for more industry partners. This not only promotes the in-depth application of AI in the field of content creation, but also sets a benchmark for technological innovation for the entire industry. In the future, with the deepening of cooperation between the two parties, I believe that it will bring more surprises to users and promote the continued prosperity of the AI content creation ecosystem.