Official on Dreamina · coming soon

Seedance 2.5 — ByteDance's 30-Second 4K AI Video Model

ByteDance presents Seedance 2.5 on an official Dreamina page, currently marked coming soon. It is not integrated on SoraVideo; the generator on this page is clearly labeled and runs Seedance 2.0.

Try Seedance 2.0 NowView Pricing

Powered by Seedance 2.0 — this form does not generate Seedance 2.5 output.

0/5000

Be specific about motion, scene, and style for the best results

Supports JPEG, PNG, WEBP, JPG, max 10MB, up to 9 files

300–6000px per side, aspect 0.4–2.5

Optional

Reference Images

Upload one or more reference files to guide generation

No reference files uploaded yet.

Supports MP4, QUICKTIME, X-MATROSKA, max 10MB, up to 3 files

Duration 2–15s per clip

Optional

Reference Videos

Upload one or more reference files to guide generation

No reference files uploaded yet.

Supports MPEG, WAV, X-WAV, AAC, MP4, OGG, max 10MB, up to 3 files

Duration 2–15s per clip

Optional

Reference Audio

Upload one or more reference files to guide generation

No reference files uploaded yet.

Seedance jobs can take a while

This model usually takes about 10-15 min. You can start up to 3 tasks at once, so you do not need to keep this page open.

If you close this page, you can come back here later or check:
Credits Required: 275|Balance: 0
Generate up to 3 videos at once — keep creating while they render.

Seedance 2.0 Demo

Configure your settings and click Generate to create an AI video with Seedance 2.0

Recent

Max 20

Sign in to save history

Your generations will sync across all devices

iFree: 1 hour • Pro: 7 days

Try Free Online

The Seedance 2.5 guide is free to read online, and official access runs through ByteDance's Dreamina product. On SoraVideo, the working form below uses Seedance 2.0 and shows its model and credit requirement before submission. The separate one-time Sora 2 Lite trial is not a 2.5 trial.

Try Current GeneratorOpen Official Dreamina

Pricing & Credits

SoraVideo pricing is credit-based: generation uses the credit amount displayed in the form, and new accounts do not receive general-purpose signup credits. Compare plan pricing on the pricing page before buying, and note that Dreamina controls its own Seedance 2.5 access and pricing.

View SoraVideo Pricing

What is Seedance 2.5?

Seedance 2.5 is the next-generation AI video model from ByteDance, unveiled on June 23, 2026 at the Volcano Engine FORCE conference. It generates a single, continuous 30-second video from one prompt — no stitching of shorter clips — while delivering native 4K output and cinematic motion.

The headline upgrade is length and resolution. Where most AI video models still stop around 15 seconds, this one produces a native 30-second clip — Dreamina also advertises a beta long-video mode up to 180 seconds — and renders directly in 4K. It also accepts up to 50 multimodal references, spanning images, video, and audio, so you can lock characters, scenes, and style across the whole shot.

Seedance 2.5 has an official product page on Dreamina, where access is currently marked coming soon, and it is not integrated on SoraVideo. Until a provider route is integrated and tested, this page lets you create videos with Seedance 2.0 and Seedance 2.0 Fast without mislabeling the output.

Seedance 2.5 generating a continuous 30-second 4K cinematic AI video from a single prompt

Key Features of Seedance 2.5

The capabilities ByteDance announced for Seedance 2.5, the new flagship AI video model

30-Second Native Clips

Seedance 2.5 generates a single, continuous 30-second video from one prompt with no stitching. Breaking past the ~15-second ceiling common elsewhere makes longer storytelling, ads, and explainer scenes possible in one generation.

Native 4K Resolution

The model renders directly in 4K, producing crisp, broadcast-ready detail with no separate upscale step for high-resolution AI video output.

Up to 50 Multimodal References

Up to 50 image, video, and audio references can be attached at once, locking character identity, scene composition, and visual style across the full 30-second clip.

Localized Editing & 3D Blockout

Localized editing and 3D blockout input give creators finer control over composition and revisions — built for precise, iterative production rather than one-shot prompting alone.

A Major Leap Over Seedance 2.0

Early demos show a significant step above Seedance 2.0 in quality, length, and realism. While the rollout continues, the same SoraVideo generator already gives you Seedance 2.0 and Seedance 2.0 Fast.
Seedance 2.5 merging up to 50 multimodal references into one cinematic AI video frame

How to Generate AI Video Now (While Seedance 2.5 Rolls Out)

The new model is not integrated on SoraVideo. Create AI video with Seedance 2.0 in four steps while keeping the model label explicit.

1

Open the Generator

Open the video generator on this page. Create a SoraVideo account, then purchase credits or choose a plan to use the current Seedance model.
2

Enter a Prompt or Image

Type a detailed text prompt for text-to-video, or upload a reference image for image-to-video. Describe the scene, motion, and style you want — the same prompting habits carry over when the 2.5 integration arrives.
3

Choose Your Settings

Pick aspect ratio, duration, and resolution. The current Seedance 2.0 model renders up to 1080p and 15 seconds here; official access to the new model and its settings remains on Dreamina.
4

Generate and Download

Generate your video, preview the result, and download it in high quality. The output remains labeled Seedance 2.0 unless a future 2.5 integration is completed and tested.
Seedance 2.5 AI video generation workflow, from prompt to finished cinematic clip

Seedance 2.5 vs Seedance 2.0 vs Kling 2.5 vs Wan 2.5

How the announced 2.5 specs compare with today's Seedance 2.0 and rival models Kling 2.5 and Wan 2.5

FeatureSeedance 2.5Seedance 2.0Kling 2.5Wan 2.5
Max Duration30s native~15s~10s~10s
Max ResolutionNative 4K1080p1080p1080p
Reference InputsUp to 50MultipleFewFew
Localized EditingLimitedLimitedPartial
Single-Prompt Long Clip———
Public AvailabilityOfficial page (coming soon)✓ Available
On SoraVideoNot integrated✓ AvailableOn Kling page—

Comparison refreshed in August 2026. The 2.5 model has an official Dreamina page but is not the model running in SoraVideo's embedded generator.

What You'll Be Able to Do with Seedance 2.5

Use cases unlocked by 30-second, 4K, multi-reference generation

Seedance 2.5 use cases — cinematic, social and 3D-animated AI video scenes

Longer Social & Ad Content

With 30-second native clips, full TikTok, Reels, and YouTube Shorts stories fit in one generation. Until the new model reaches SoraVideo, you can already build shorter social videos with Seedance 2.0.

Cinematic & 3D Animation

4K output and 50-reference control suit Pixar-style 3D animation, cinematic shorts, and character-consistent scenes — the look creators are already testing in early demos.

E-commerce & Marketing

Native 4K and longer runtime are strong for product videos and marketing assets. The current Seedance 2.0 model on SoraVideo covers these e-commerce and ad workflows today.

Seedance 2.5 — Release Date & Availability

The model now has an official product page on Dreamina, where access is marked coming soon. Availability there is separate from SoraVideo, where it is not integrated.

  • Seedance 2.5 was developed by ByteDance and unveiled at the Volcano Engine FORCE conference on June 23, 2026
  • It is not currently available in the SoraVideo generator
  • Announced specs: native 30-second clips, native 4K output, up to 50 multimodal references, and a beta long-video mode up to 180 seconds
  • ByteDance presents the model on an official Dreamina page, with access marked coming soon
  • SoraVideo offers Seedance 2.0 and Seedance 2.0 Fast so you can create AI video now
  • This page will change its generator label only after a 2.5 integration is completed and tested

For current access status and pricing, refer to the official Dreamina product page.

SoraVideo has not added a Seedance 2.5 API route in this release. A future selector will require an official provider path plus verified request, credit, task-status, and delivered-output behavior. The embedded generator continues to use the existing Seedance 2.0 integration.

Visit Official Dreamina

Seedance 2.5 FAQ

Common questions about Seedance 2.5, ByteDance's new AI video model

Is Seedance 2.5 out yet?

ByteDance runs an official Seedance 2.5 page on Dreamina, where access is currently marked coming soon. It is not integrated on SoraVideo; this page's generator runs Seedance 2.0.

What's new in Seedance 2.5 vs Seedance 2.0?

The 2.5 release adds single-prompt 30-second native clips, native 4K output, up to 50 multimodal references, and localized editing. Seedance 2.0 currently tops out around 15 seconds and 1080p.

Can it really generate 30-second videos?

Yes. ByteDance says a single continuous 30-second clip comes from one prompt without stitching — and Dreamina advertises a beta long-video mode up to 180 seconds.

How many references does Seedance 2.5 support?

Up to 50 multimodal references — images, video, and audio — to lock characters, scenes, and style across the clip.

Does it support 4K?

Yes, rendering is natively 4K, so no separate upscaling step is needed for high-resolution output.

When will Seedance 2.5 reach SoraVideo?

There is no confirmed integration date. The official Dreamina page is the source for its own access timing; this page updates its generator label only after the provider route is integrated and tested.

Who makes it?

Seedance 2.5 is built by ByteDance and was announced at its Volcano Engine FORCE conference on June 23, 2026.

Where can I use it now?

Follow ByteDance's official Dreamina product for upcoming access. On SoraVideo you can generate today with Seedance 2.0 and Seedance 2.0 Fast; those outputs are not labeled as 2.5.

Compare Related Video Models

Review distinct model pages and their current generator disclosures.

MiniMax H3 TrackerRead the official H3 release guide and generate with Hailuo 2.3.Hailuo 2.3Open MiniMax's current cinematic video workflow.Kling 3.0Compare a separate released AI video model workflow.

Create AI Video with the Current Seedance Model

Review Seedance 2.5 on official Dreamina, or generate with clearly labeled Seedance 2.0 on SoraVideo today.

Try Seedance 2.0View Pricing Plans

Seedance 2.5 is a product developed by ByteDance. SoraVideo is independent and is not affiliated with ByteDance. Release information was refreshed from the official Dreamina product page in August 2026. The generator on this page currently runs Seedance 2.0, not the 2.5 model.