30-second native clips
A single continuous 30s shot, generated natively — no stitching, no extension passes.
Preview
Seedance 2.5
ByteDance’s next video model — 30-second native clips and up to 50 multimodal references. Expected to launch in early July 2026.
What is Seedance 2.5?
Seedance 2.5 is ByteDance’s next-generation video model, previewed at the June 2026 Force conference. It generates a single native 30-second clip in one pass and accepts up to 50 multimodal reference materials — images, video and audio — for much stronger control than Seedance 2.0. A public launch is expected in early July 2026.
Not yet public — run Seedance 2.0 in Renoise today. Seedance 2.0
Seedance 2.5: what ByteDance announced
Previewed at the Force conference on June 23, 2026, alongside its first creative short film. Seedance 2.5 looks powerful — a clear step up from Seedance 2.0.
Up to 50 references
Feed up to 50 full-modal reference materials — images, video and audio — to steer a single generation.
4K resolution
Seedance 2.0 already supports 4K, and Seedance 2.5 is expected to support 4K as well.
Early-July launch
Reports point to a public launch in early July 2026. It’s in late closed-beta now — Renoise integrates new models the moment they ship.
What can creators do with Seedance 2.5?
The headline change is length. Most AI video tools cap out at 15–20 seconds, forcing you to stitch clips together — which invites character drift and continuity breaks. Seedance 2.5 generates a single, continuous 30-second shot natively, so a complete scene, ad or story beat holds from the first frame to the last in one pass.
The second is input capacity: up to 50 full-modal reference materials in a single generation. Feeding images, video and audio together lets the model lock onto brand elements and complex instructions — exactly what brand films and episodic short dramas demand. The third is control. Flexible, localized editing fixes a small flaw without regenerating the whole clip, and 3D blockout (白模) input lets directors pre-stage camera and composition for frame-level command over the result.
Taken together, that moves AI video from short, single-shot novelty toward production-grade work: longer narratives, consistent characters and the kind of precise, repeatable control professional pipelines need.
Seedance 2.5 vs. Seedance 2.0
ByteDance’s announced Seedance 2.5 capabilities next to Seedance 2.0, the model Renoise runs today.
| Capability | Seedance 2.5 | Seedance 2.0 |
|---|---|---|
| Max clip length | 30s native, one shot | 4–15s |
| Reference inputs | Up to 50 multimodal | 9 images + 3 video + 3 audio |
| Localized editing | Edit regions, no full re-render | — |
| 3D blockout input | Yes | — |
| Resolution | Expected 4K | Native 4K |
| Availability | Expected early July 2026 | Live in Renoise |
Don’t wait for 2.5. Start with Seedance 2.0.
Generate today on Seedance 2.0, then upgrade your shots to Seedance 2.5 in Renoise the moment it launches.
FAQ
1.Is Seedance 2.5 out yet?
Not for public use. ByteDance previewed Seedance 2.5 at its Force conference on June 23, 2026, and it’s in late closed-beta with a global enterprise beta now recruiting. Public reports point to an early-July 2026 launch, but no firm date is officially confirmed.
2.What’s new in Seedance 2.5?
ByteDance’s announcement highlights three things: a single 30-second native clip generated directly (no stitching), support for up to 50 full-modal reference materials in one generation, and more controllable video generation and editing. It builds on Seedance 2.0’s unified multimodal audio-video architecture.
3.Can Seedance 2.5 generate 30-second videos?
Yes. ByteDance says Seedance 2.5 produces a single, continuous 30-second clip natively — no stitching or extension passes — which keeps characters and scenes consistent across a longer narrative than the 15–20 seconds most tools allow.
4.How many reference materials does Seedance 2.5 support?
Up to 50 full-modal reference materials in one generation, combining images, video and audio. That capacity is aimed at brand films and episodic short dramas, where the model has to hold consistent brand elements and follow complex, multi-part instructions.
5.Does Seedance 2.5 support video editing and 3D blockout?
Yes. ByteDance highlights flexible, localized editing — you can change a region of a clip without regenerating the whole video — plus 3D blockout (白模) input, letting you pre-stage camera and composition for tighter control over the final frame.
6.Does Seedance 2.5 support 4K?
Seedance 2.5 is expected to support 4K output, and its predecessor Seedance 2.0 has already been upgraded to native 4K. ByteDance’s confirmed headline specs are the 30-second native clip and up to 50 multimodal references.
7.When will Seedance 2.5 release?
No official launch date has been published. Coverage points to early July 2026, and it’s already in a late closed-beta with a global enterprise beta recruiting. Official pricing has not been announced yet.
8.Who makes Seedance 2.5?
Seedance is developed by ByteDance, under Volcano Engine (豆包). It was unveiled at the Force conference on June 23, 2026, alongside its first creative short film.
By Max, RenoiseLast reviewed Models verified: Seedance 2.0