Seedance2.5, a large model of bean bag video generation, will be released in early July: it supports 30-second video and 50 full-modal material input

On June 23, Seedance2.5, a new video generation model for bean buns, is currently at the end of internal testing and will be officially released in early July. It is reported that this new model has mainly achieved three core technology upgrades: breakthrough in generation time: single video generation time can reach up to 30 seconds, greatly improving lens coherence; extreme multimodal input: "multi-reference capabilities" have been greatly upgraded, Support the joint input of up to 50 all-modal materials (claimed to be the largest in the world), allowing a large number of image assets to be input at one time for editing; improved editing control: It has more flexible and controllable video editing and generation capabilities, aiming to further improve creator efficiency and finished product quality. (Hard AI)