01 / 480p
12credits / second
Without video input
For text, image, frame-pair, and supported audio-guided workflows without a reference video.
ByteDance flagship multimodal video modelAI Motion System / 02.5
Now available on Toono AI
Direct every frame with Seedance 2.5
Move from a written scene to campaign-ready motion with native audio, up to 30 seconds of output, start and end frame control, and a much larger multimodal reference canvas.
30s
Native long-form output
30+10+10
Image, video, and audio references
12 cr/s
Entry price at 480p
Official Replicate sample
5 seconds · 720p · native audioSource prompt
a golden retriever puppy running across a green meadow toward the camera, slow motion, warm afternoon light
View sourceFrame 00361280 × 720 / H.264
A video model built for direction, not just prompting
Choose the input language that best matches the shot: a scene brief, a keyframe pair, or a multimodal reference set spanning images, motion, and sound.
Text-to-video
Describe subject behavior, camera language, lighting, and edit rhythm in one production brief.
Start and end frames
Anchor the opening image and optionally define the final frame for a more deliberate transition.
Up to 30 reference images
Build a broad visual board for characters, products, wardrobe, environments, and art direction.
Up to 10 reference videos
Carry movement, camera behavior, and pacing into a new shot with motion references.
Native audio and references
Generate synchronized sound and guide it with up to 10 audio references when visual references are present.
Flexible output framing
Create in 480p or 720p across landscape, square, portrait, ultrawide, and adaptive aspect ratios.
Seedance 2.5 Generator
No Videos Generated
Transparent usage pricing
Choose the input path. Pay by output second.
Seedance 2.5 uses separate rates for generations with and without video input. The displayed credit estimate updates before you submit, and server-side calculation remains the source of truth.
02 / 480p
50credits / second
With video input
For workflows that include one or more reference videos.
03 / 720p
24credits / second
Without video input
Higher-detail output without a reference video in the input set.
04 / 720p
100credits / second
With video input
Higher-detail output for video-referenced generation and editing workflows.
A 5-second clip costs 60, 250, 120, or 500 credits respectively. Final cost depends on resolution, duration, and whether video input is attached.
Compare credit plansProduction workflow
From scene direction to a finished motion asset
Use the simplest input set that communicates the shot, then refine from the result instead of rebuilding the brief every time.
1
Write the scene
Define the subject, action, camera, environment, lighting, and audio intent.
2
Choose references
Use a start/end frame route or a multimodal reference set; the two input families remain separate.
3
Set output controls
Select 5 to 30 seconds, 480p or 720p, aspect ratio, and native audio.
4
Generate and iterate
Review the result, reuse the strongest take, and tighten direction for campaign variants.
Seedance 2.5 questions, answered
Know the input rules and credit path before starting a longer generation.
Your next campaign shot can start here
Test Seedance 2.5 on this page, then move into the production workspace for generation history and repeated iteration.