ByteDance officially launched Seedance 2.5 on July 31. The Seed team says the model is rolling out on Jimeng AI and Doubao Pro, with API access planned for Volcano Engine ModelArk. Caixin and Sina Finance reported the same rollout.
Thirty seconds changes the production unit
“The model can organize multiple logically connected shots.”
Seedance 2.5 lifts single-pass generation from 15 seconds to 30 seconds and supports multi-round extensions. Thirty seconds is close to the minimum unit for an ad, a classroom clip or a short-drama scene.
References become production input
The model accepts up to 30 images, 10 videos and 10 audio clips in one pass. Dreamina summarizes this as up to 50 multimodal references. Product photos, characters, music direction and sample footage can move into one workflow.
Editing is the practical test
Timestamp-level editing is the most useful part of the release. Teams need to fix a product placement, camera move or action segment without regenerating the whole clip. ByteDance also says complex physical motion and crowded multi-subject scenes still need improvement.
Caixin reported that Seedance 2.0's overseas call share rose from one-third to one-half by the end of June. The next checks are API timing, pricing, concurrency limits and third-party tests of multi-reference and timestamp editing.
Fuentes: ByteDance Seed official release, Caixin, CocoLoop, Sina Finance; Seedance project page and Dreamina verified 30-second generation, 30-image/10-video/10-audio references, timestamp editing, Jimeng AI/Doubao Pro availability and Volcano Engine ModelArk API scope.