Seedance 2.5
The full model page: what the 30-second window changes, how references behave, and worked prompt examples.
Read more: Seedance 2.5Seedance 3.0 Guide:AI Video Generator
Official models, no watermark, up to 1080P — generate online free, with credits at sign-up and no credit card required.
Coming Soon
What Is Seedance 3.0?
Seedance 3.0 is the upcoming Seedance AI video model. It has not been released yet; official capabilities and specifications will be updated after launch.
| Clip length | Coming Soon |
|---|---|
| Resolution | Coming Soon |
| Audio | Coming Soon |
| Inputs | Coming Soon |
| Editing | Coming Soon |
| Watermark | Coming Soon |
Last checked: 2026-10-23
Seedance 2.0
The previous generation, still the cheaper pick for short 4–15 second shots with multimodal references.
Read more: Seedance 2.0Wan 3.0
Alibaba's 30-second model with native audio, already live here: the closest like-for-like comparison while Seedance 3.0 is unreleased.
Read more: Wan 3.0Text to video
Open the generator with a blank prompt and start from a description rather than a reference image.
Read more: Text to videoQuestions
FAQs about Seedance 3.0
No. As of October 23, 2026, ByteDance has not announced or released a model called Seedance 3.0. There is no official model page and no API model name for it.
There is no announced date. ByteDance has not given a release window, opened a waitlist or published a teaser, so any specific date you see elsewhere is a guess rather than an announcement. The moment an official date exists, this page will be updated.
Seedance 2.5. It generates up to 30 seconds at up to 1080P with audio produced in the same pass, accepts image and video references, and can edit or extend an existing clip. You can run it free from the composer at the top of this page.
If ByteDance releases Seedance 3.0, we will bring it online here right away.
Available now
Waiting for 3.0? Start with 2.5.
Thirty seconds, 1080P, generated audio and no watermark — free to try, no card required.
Try Seedance 2.5 free