Home/Models/Seedance 2.5
ByteDancevideo generation

Seedance 2.5

30-second clips, 30 reference images, and the best character consistency available

Seedance 2.5 is the current flagship of ByteDance's video line and the default model in the Framiq studio. It doubles the clip length of Seedance 2.0 to 30 seconds and raises the reference limit from 9 images to 30 — which is what makes reliable character consistency possible, since the model sees a face from far more angles before it renders. It also accepts an existing video as a reference, so a clip can be extended or restyled rather than regenerated from scratch. Combined with Framiq's Extra Long Video mode, which chains clips and stitches them, it produces continuous sequences of up to five minutes.

Resolution

480p / 720p

Duration

4–30s

Image refs

≤30 images

Video refs

Yes — extend & edit

Key features

  • Up to 30s per clip
  • Up to 30 reference images
  • Best-in-class character consistency
  • Video-to-video editing
  • AI audio generation
  • Start & end frame control

Use cases

  • Recurring AI characters
  • UGC-style ad creative
  • Long-form narrative video
  • Product demos with a consistent presenter
  • Social series with the same face

Model versions

Seedance 2.5

Try Seedance 2.5 now

Start with 20 free credits. No credit card required.

Get started free

Other AI models on Framiq