On February 18th, LeapStar and Geely Automobile Group jointly announced the open sourcing of two multi-modal large models from their jointly developed LeapStep series to global developers. These models encompass the world's largest and most efficient open-source video generation model, LeapStep-Video-T2V, and the industry's pioneering product-level open-source speech interaction model, LeapStep-Audio. Effective immediately, users can experience these cutting-edge technologies within the Yuewen App. As a strategic partner within Geely Automobile's technology ecosystem, LeapStar has deepened its collaboration with Geely in areas such as computing power algorithms and scenario training, resulting in significant enhancements to model performance. This open-source endeavor aims to foster sharing and innovation in large model technology, thereby driving the inclusive advancement of artificial intelligence.