Vidu Q2 is a multimodal video generation model for creators, focusing on generating high-quality short films and long clips. It can integrate multiple references in a single generation process, and even supports uploading multiple references at once, locking in the appearance of people/objects/scenes and maintaining consistency across shots. It supports long-duration/segment continuation and splicing workflows, and improves the efficiency of long videos based on cascading/parallel inference.