Seedance is ByteDance's video generation model family, built by its Seed research team. The current model, Seedance 2.5, generates video and audio together from text, images, video and audio references, with clips of up to 30 seconds that can be extended twice; Seedance 2.0 and its cheaper Fast and Mini versions remain available. It suits developers who want a video model with native audio behind an API, and creators who already work in ByteDance's Dreamina or CapCut apps.
The core caveat is that Seedance has no subscription of its own. On the BytePlus ModelArk API it is billed by tokens, which grow with clip length, resolution and frame rate, so the price of a clip depends on its settings; failed generations are not charged. In apps it is paid for with the app's credits, and the credit cost per clip is set by each app.
Developers buy access through BytePlus ModelArk. Creators can use it inside Dreamina's monthly credit plans or CapCut's AI credits instead, each billed separately from the API.