ByteDance unveils Seedance 2.5 with native 30-second AI video

Read the original article →

What happened

ByteDance announced Seedance 2.5 on June 23 at its Volcano Engine FORCE conference. The model powers Dreamina, Doubao, and CapCut, and the company claims a single native 30-second clip without stitching, up to 50 reference inputs per generation, and inpaint-style frame editing.

Seedance 2.0 also received a native 4K upgrade and leads independent blind-preference testing for AI video at lower cost. Seedance 2.5 is framed as a global enterprise beta, with a public launch targeted for early July. It competes with OpenAI's Sora, Google's Veo, and Runway.

Why it matters

Clip length and reference control are the practical limits that have kept AI video out of many production workflows. A native 30-second generation, if it holds up, reduces the stitching and continuity work that eats editing time.

Strong blind-preference results at lower cost also pressure competitors on price, not just quality.

MintedBrain take

Headline durations and reference counts are best verified on your own footage before you plan around them. If the 30-second and 50-reference claims survive hands-on testing, Seedance becomes a serious option for short-form production.

References

This article was originally published at TechTimes. For the full piece, read the original article.

Discussion

  • Loading…

← Back to News