Vchitect 2.0 (Dream Builder 2.0) is a high-end video generation model developed by the Shanghai Artificial Intelligence Laboratory, designed to empower new possibilities in video creation. It supports 20-second video generation, flexible aspect ratios, enhanced spatial-temporal generation, and long video assessments. Vchitect 2.0 utilizes advanced technology to transform static images into dynamic 5-10 second videos, allowing users to easily convert photos or designs into captivating visual experiences. Furthermore, Vchitect 2.0 facilitates the evaluation of long video generation models through the VBench platform, providing comprehensive and continuously updated evaluation rankings that support various long video models like Gen-3, Kling, OpenSora, and more.