30s Continuous Video
Create complete cinematic sequences without stitching clips together.
ByteDance’s June 2026 preview: native 30-second clips, up to 50 multimodal inputs, stronger control.
Dreamina’s public Seedance 2.5 page frames the model as a controlled production workflow, not just a longer text-to-video generator.
Create complete cinematic sequences without stitching clips together.
Combine images, videos, audio, and style references in a single generation.
Edit specific elements while preserving the original scene, motion, and style.
Generate natural voices and synchronized lip movements across multiple languages.
Seedance 2.5 is framed around controllable production workflows: references, multimodal inputs, precision editing, stable 30-second generation and multilingual creation.
Use green-screen, white-model or motion references to guide character movement, spatial layout and interactions. The goal is stronger scene control than text prompts alone.
Create a native 30-second clip in one pass while preserving identity, lighting, camera movement and scene continuity across the full shot.
Combine prompts, scripts, reference photos, videos, music and style guides in one generation so the model can follow a complete creative direction.
Refine only the parts of a video that need adjustment while preserving lighting, composition, motion, audio and timeline continuity.
Create across major languages including Chinese, English, Spanish, Arabic, Portuguese, Vietnamese, Japanese and Korean, with cleaner 4K output for editing.
Dreamina positions Seedance 2.5 around publishable commercial and cinematic work: social content, product ads, brand campaigns and short-form storytelling.
Seedance 2.5 is coming soon. ByteDance previewed the model at the June 23, 2026 Force conference, and BytePlus has announced the Seedance 2.5 API launch for July 2026.
In Renoise, Seedance 2.0 is live today on Canvas. We’ll publish a hands-on Seedance 2.5 review as soon as the API goes live and integrate it as soon as it’s available, so teams can keep producing now and move shots over when it lands.
June 23, 2026
ByteDance / Volcano Engine previewed Seedance 2.5 at the Force 原动力大会 and released the model’s first creative short film.
Coming soon
BytePlus has announced Dreamina Seedance 2.5 as coming soon, with preview demos and capabilities shared ahead of launch.
July 2026
The Seedance 2.5 API is scheduled to launch on BytePlus in July 2026. Renoise runs Seedance 2.0 today and will integrate 2.5 as soon as it is available.
Public Seedance 2.5 capabilities next to Seedance 2.0, the model Renoise runs today.
| Capability | Seedance 2.5 (Recommended) | Seedance 2.0 |
|---|---|---|
| Max clip length | 30s native, one shot | 4–15s |
| Asset inputs | Up to 50 assets | 9 images + 3 video + 3 audio |
| Precision editing | Edit regions, no full re-render | — |
| 3D blockout input | Yes | — |
| Resolution | Expected 4K | Native 4K |
| Availability | Coming soon — July 2026 on BytePlus | Live in Renoise |
Use Seedance 2.0 now. We’ll review Seedance 2.5 and add it when it becomes available.
One Renoise plan unlocks Seedance 2.0 today, plus every available video and image model.
Not in Renoise yet. ByteDance previewed Seedance 2.5 at its Force conference on June 23, 2026, Dreamina describes it as coming soon, and BytePlus has announced the API launch for July 2026. Until then, Renoise runs Seedance 2.0.
The big changes are control and duration: 30-second standard videos, R2V references, up to 50 multimodal inputs, precision region editing, 4K output and cleaner multilingual creation.
R2V refers to reference-to-video control. Dreamina describes green-screen and white-model references that can guide motion, spatial layout and character interactions, so the generation follows a structured visual plan instead of relying only on text.
Yes. Seedance 2.5 is positioned around 30-second standard generation, which helps creators produce a full scene, ad beat or social cut with fewer stitched clips.
Up to 50 multimodal inputs in one generation. That can include prompts, scripts, reference photos, videos, audio, music and style guides, which is useful for brand work where the model needs to follow more than a single text prompt.
Dreamina highlights local video editing, where selected regions or objects can be refined without recreating the whole scene. It also describes white-model / blockout-style references for controlling motion and spatial composition.
Dreamina’s public Seedance 2.5 page describes cinematic 4K video output. Seedance 2.0 already supports native 4K in Renoise today.
Seedance 2.5 is positioned for social media videos, product ads, ecommerce showcases, brand campaigns and cinematic short stories — workflows that benefit from longer scenes, stable characters, R2V motion guidance and local fixes.
The Seedance 2.5 API is expected to launch in July 2026 on BytePlus, which has officially announced the model as coming soon. Renoise will publish a hands-on review and integrate Seedance 2.5 as soon as it’s available.
Public pricing for Seedance 2.5 has not been announced yet — the API is expected to launch on BytePlus in July 2026. Renoise runs Seedance 2.0 today on credit-based plans and will add Seedance 2.5 as soon as it’s available, so you can start generating now and move shots over when it lands.
Not yet — Seedance 2.5 is coming soon, with the API expected to launch in July 2026 on BytePlus. Renoise runs Seedance 2.0 today on the Canvas, and we’ll publish a hands-on Seedance 2.5 review and integrate it as soon as it’s available.
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.