[Zhongtai Media & Internet] ByteDance Releases Next-Generation Video Generation Model Seedance 2.5

English summary

ByteDance releases next-generation video generation model Seedance 2.5 [Zhongtai Media & Internet] ByteDance released Seedance 2.5, continuing the multimodal audio-video joint generation architecture, with a focus on breakthroughs in long-form narrative, multimodal reference, and video editing capabilities. The model supports single-generation of up to 30 seconds of video and supports multiple rounds of extension, maintaining consistency of characters, scenes, shots, and audio-visual elements in long videos, achieving more complete story expression. At the same time, the model further upgrades multimodal reference capabilities, supporting up to 30 images, 10 videos, and 10 audio clips as input, enabling comprehensive understanding of characters, actions, styles, scenes, etc., for more complex and controllable video creation. Compared with previous AI video models that focused mainly on image quality and single-shot generation, Seedance 2.5 places greater emphasis on the integrity of content production workflows and commercialization capabilities. …