ByteDance’s Seedance 2.5: 30-second clips and targeted edits

Models
Video
The video model behind Dreamina doubles clip length and lets you change one moment or one person in a video.
Published

July 31, 2026

What happened

  • On July 31, ByteDance made Seedance 2.5 the default video model in its Dreamina and Jimeng apps, after a preview in June. Developers can use it through BytePlus.
  • It makes one continuous clip of up to 30 seconds (Seedance 2.0 stopped at 15) and can use up to 50 reference pictures, clips or sounds to keep people and places looking the same.
  • It adds targeted editing: change a single moment, character or object in a video without making it again.
  • As of mid-October it is second on the Arena video editing board, just behind Wan 3.0.

Why it matters for you

Longer single clips mean less stitching and fewer jumps between scenes. Targeted editing is what home-video editors need most: fix one shot, keep the rest. Dreamina is ByteDance’s consumer app, from the makers of CapCut, so it is one of the easiest ways to try a top video model.

Sources

Notes for learning. Nothing here is investment advice. All news