ByteDance Launches Seedance 2.5 With 30-Second Native Clips and Maya, Blender Plugins
AI video model releases
ByteDance Launches Seedance 2.5 With 30-Second Native Clips and Maya, Blender Plugins
ByteDance globally launched Seedance 2.5 through Dreamina, a text-to-video and reference-to-video model available to subscriber accounts.
The model natively generates up to 30 seconds in a single pass at up to 4K resolution, with a beta long video mode extending output to 3 minutes.
Seedance 2.5 accepts as many as 50 multimodal references per clip and supports localized editing of specific regions and timestamps without regenerating the full shot.
ByteDance shipped plugins for Autodesk Maya and Blender, aiming to integrate the model directly into existing 3D production and previsualization pipelines rather than requiring frame export and browser upload.
Most competing video models still cap generations at 5 to 10 seconds, making Seedance 2.5's longer coherent takes a direct attack on the stitching problem that has limited professional adoption.
cho.sh