Video model
Choose a Seedance workflow, tier and source set before generation
Seedance is a formal video model in WAN3 with text, frame-guided and reference-led workflows. Compare its verified 720p controls, three service tiers, source limits and credit calculation before submitting.
01
Start from text, frames or mixed references
Text-to-video starts from a prompt. Image-to-video requires a first frame and can accept an optional end frame. Reference-to-video accepts one to twelve references, with up to nine images, three videos and three audio files inside that total.
- All launch workflows support durations from 4–15 seconds.
- Supported launch output is 720p with square, cinematic, landscape and portrait aspect ratios.
- Optional generated audio is available from the advanced controls.
02
Select Mini, Standard or Pro with a visible quote
Mini is priced at 17.5 credits per second, Standard at 28 credits per second and Pro at 35 credits per second. The final task price is rounded to the nearest five credits and displayed before submission.
03
Prepare source files within the launch limits
Frame images accept JPEG, PNG or WebP up to 30 MB. Reference video and audio files accept the formats listed by the Workspace up to 50 MB each. Unsupported media is rejected before generation.
04
Track private status, refund and delivery
Seedance tasks use owner-only queued, processing, succeeded, failed and refunded states. A successful result is copied to WAN3-managed storage before the download becomes available.