Thirty seconds, one take.
Double the ceiling of 2.0. Up to 30 seconds in a single generation, with native audio and thirty reference images, so a whole scene comes out of one prompt instead of four stitched clips.
From 58 credits per 5 seconds. Video generation starts on the paid plans. Native output is 480p and 720p.
ByteDance
Seedance 2.5
What is Seedance 2.5?
Seedance 2.5 is ByteDance's long form video model, launched in July 2026. It generates a single continuous clip of 4 to 30 seconds with a synchronized audio track, and accepts up to 30 reference images, 10 reference videos and 10 audio inputs in one request. That reference budget is the point: a cast, a product line and a location can all be pinned at once, so a thirty second scene stays consistent without the character drift that comes from stitching four separate generations together. On APImage it renders natively at 480p or 720p and draws on the same credit pool as every other model.
- Provider
- ByteDance, via BytePlus ModelArk
- Resolution
- 480p, 720p, native
- Frame rate
- 24fps
- Duration
- 4 to 30 seconds in one generation
- Audio
- Native, generated with the picture
- Modes
- Text to video, image to video, video to video
- Aspect ratios
- 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
- Reference images
- Up to 30 per generation
- Reference videos
- Up to 10 per generation
- Audio inputs
- Up to 10 per generation
- Cost
- From 58 credits per 5 seconds
What Seedance 2.5 does that the others do not.
Thirty seconds, uncut
A full clip in one pass, with no stitching and no seam where two generations were taped together. Twice the ceiling of Seedance 2.0.
Thirty references
Images, clips and audio, tagged and addressable from the prompt. Enough to lock a cast, a product line and a location in the same generation.
Audio in the box
Speech, effects and ambience are generated with the picture and land on the beat, so there is no separate pass to sync afterwards.
Identity that survives the take
Faces and products hold across thirty seconds. Length is where consistency usually breaks, and this is the model built for it.
Where it earns its credits in real work.
Full length ads
A thirty second spot is the standard unit of paid media. This renders one as one thing, not as a stitched sequence.
Multi shot narrative
Set up, develop and pay off inside a single generation, with the same cast in every beat.
Explainers and demos
Long enough to walk through a product properly, with narration generated alongside the picture.
Character driven UGC
Thirty reference images pin a presenter hard enough to reuse them across a whole campaign.
Music led social
Native audio means the cut lands on the beat instead of being nudged into place in an editor.
- 1Open the video studio and pick Seedance 2.5.
- 2Upload your references. Up to 30 images, 10 clips and 10 audio files, tagged so the prompt can address them.
- 3Write the scene, including what you want to hear as well as what you want to see.
- 4Set the length anywhere up to 30 seconds and generate.
Seedance 2.5 is not the right answer to everything. Your plan includes all of them, so switching model costs nothing.
FLUX 3 Video
You need Full HD. Seedance 2.5 renders natively at 720p.
See FLUX 3 VideoSeedance 2.0
You want native 1080p and the clip fits in 15 seconds.
See Seedance 2.0FLUX 3 Video Draft
You are still exploring and want the cheapest possible look at an idea.
What it costs
From 58 credits per 5 seconds
Video generation starts on the paid plans. Native output is 480p and 720p.
No separate subscription
Seedance 2.5 is on the same plan as every other model on APImage, priced in the same credits out of the same pool. Run the same prompt through three models and keep the one that looks right.
New models never cost extra
When a provider ships something new it appears in your studio at no additional charge. Ultra plans get it first.
is live on APImage
One subscription, one pool of credits, and not a single extra login. Start free and switch models with a dropdown.
Generate with Seedance 2.5