Recently, LiblibAI, China's largest AI image creation platform, announced its integration with Alibaba's Tongyi series of large language models, further enhancing its AI creation capabilities. Leveraging the newly open-sourced Wanxiang model, LiblibAI has launched text-to-video and image-to-video features, allowing users to generate 10-second short videos using text prompts or uploaded images.
Data shows that Wanxiang 2.1 (Wan2.1) has accumulated over 2 million downloads on Hugging Face and ModelScope, and boasts over 8.7k stars on GitHub, demonstrating strong market recognition.
Simultaneously, LiblibAI has also developed a prompt optimization window based on Alibaba's Qwen-turbo large language model. This feature utilizes system prompts to help users optimize their input and supports automatic translation of Chinese prompts into English prompts understandable by the model. Following the integration of Qwen-turbo, LiblibAI's prompt optimization time has reportedly been reduced by over 50%, further improving the efficiency and user experience of AI creation.
As the first AI platform in China to obtain national registration, LiblibAI has accumulated over 20 million creators and possesses more than 100,000 original models, providing AI image creation solutions for numerous enterprises and institutions, including Tmall Campus, Luckin Coffee, and Tsinghua University. With its "open-source model + modular workflow" platform model, LiblibAI has built a complete ecosystem integrating AI content creation, sharing, copyright management, and sales. Notably, the platform has completed four rounds of financing in the past year, setting a record for financing speed in China's AI application sector, showcasing strong market competitiveness and development potential.
