ByteDance has released Seedance 2.5, an AI video model that generates video and audio together rather than requiring separate production steps. The model can produce clips up to 30 seconds long and extend them multiple times.
The update is aimed at more coherent short-form production. Users can provide up to 30 images, 10 video clips, and 10 audio files as references, which can help maintain characters, camera angles, and audio cues across a scene. ByteDance also says textures, lighting, and skin detail have improved.
For ad teams and creators, the practical change is workflow. Instead of building one short silent segment and adding sound later, teams can generate a longer audiovisual scene inside one pipeline. ByteDance showcased the model with a short film made entirely in Seedance 2.5.
The model is available through Jimeng AI and Doubao Pro, with API access planned through BytePlus ModelArk. As with other generative video systems, users will still need to check rights, realism, and disclosure expectations before publishing outputs.