VideoByteDanceByteDance

Seedance 2.5

ByteDance's 30-second single-shot video model, with up to 50 multimodal references and native synchronized audio.

Max Duration
30 seconds max
Resolution
480p, 720p
Aspect Ratios
16:9, 9:16, 1:1, 4:3, 3:4, 21:9
Audio
Native audio
Up to 720p resolution30 sec max durationNative audioUp to 50 references
Seedance 2.5 30-second single-shot generation

What makes Seedance 2.5 unique

01

Native 30-second single-shot generation

one continuous take, no stitching between clips

02

Up to 50 multimodal references across images, video and audio in a single request

03

Positional reference binding

name a character in the prompt and it holds across the whole take

04

Optional first-frame and last-frame control for precise scene anchoring

05

Native synchronized audio, including ambient sound and lip-synced speech

06

Any whole duration from 4 to 30 seconds, priced per second

Best for

Long-form product and launch videosShort films and story trailersBranded storytelling with a consistent castVirtual influencer and creator contentScenes that would otherwise need multi-clip stitching

Prompt tips

  1. 1.

    Write the whole 30 seconds as one continuous shot — describe how the scene evolves rather than cutting between beats.

  2. 2.

    Attach a reference image per character and name them in the prompt; the binding holds for the full take.

  3. 3.

    Reference videos are the fastest way to pin a camera move or a motion style you already have.

  4. 4.

    Include audio cues in your prompt; sound is generated with the picture, not added after.

  5. 5.

    Start at 480p while you iterate — the per-second price at 720p is more than double.

How Seedance 2.5 works

Seedance 2.5 reasons about the whole shot at once rather than generating and stitching short clips, which is what keeps motion, lighting, and subject identity coherent from the first frame to the last across a full 30 seconds. Reference images, videos, and audio are bound by upload position, so a character named in the prompt maps to a specific reference throughout the take.

How to use on Flashloop

Step 01

Select Seedance 2.5

Open Flashloop and choose Seedance 2.5 from the model selector in the video creator.

Step 02

Write your prompt

Describe what you want to see in your video, be as detailed as you like.

Step 03

Generate & download

Hit generate and your video will be ready in seconds. Download or share directly.

Frequently asked questions

Up to 30 seconds, and any whole number of seconds from 4 upward. It is a single continuous take, not several short clips joined together.

2.5 doubles the maximum length to 30 seconds and widens references from 7 images to 30 images plus 10 videos and 10 audio clips. Seedance 2.0 still goes higher on resolution — it reaches 1080p and 4K, where 2.5 tops out at 720p.

480p and 720p. There is no 1080p or 4K tier on 2.5 — if you need those, use Seedance 2.0 on High.

Up to 30 reference images, 10 reference videos, and 10 reference audio clips per generation, with a combined ceiling of 50 files. Reference video and audio are each capped at 30 seconds in total. First-frame / last-frame control is also supported.

Yes. Sound effects, ambient audio, and lip-synced speech are generated together with the picture rather than added afterwards.

Other video models

View all →

Ready to create with Seedance 2.5?

Available on iOS, Android, and web.

Start Creating