Choose a category
3 models available
8 models available
3 models available
5 models available
4 models available
8 models available
Seedance 2.0 Fast
Seedance 2.0 quality at half the price and wait.
Creative Vision
Reference uploads as @Image1, @Image2 … in the prompt.
Video Parameters
AI audio synced to the video. Does not change the price.
Cinema Preview
Generation History0 records
Seedance 2.0 Fast AI Video Generator
Seedance 2.0 Fast is the middle tier of ByteDance's Seedance 2.0 line: the same @Image reference casting and the same controls as the full model, capped at 720p, at roughly 60% of the price and a shorter wait. It is the tier to iterate on.
Prompts are portable across all three Seedance tiers — whatever you settle on here runs unchanged on Standard
Why Choose Seedance 2.0 Fast
ByteDance ships Seedance 2.0 in three tiers that share one architecture and one parameter set. Fast is not a cut-down model with different behaviour — it is the same model rendering at a lower ceiling, which is exactly what you want while a prompt is still changing.
Shorter Wait per Attempt
Fast returns sooner than the standard tier, and on a page where you may run a prompt six times before it lands, the minutes saved per attempt add up faster than the credits do.
Roughly 60% of Standard's Price
Five seconds at 720p is 30 credits here against 50 on Standard. Ten seconds is 60 against 95. The same budget buys you noticeably more attempts at the same resolution.
Full @Image Reference Casting
Up to five reference images, addressed as @Image1 through @Image5 in the prompt, with the same role-based assignment as the full model. None of the direction is held back for the higher tier.
Prompts Transfer Upward
Identical parameters across tiers means a prompt developed here runs on Standard with nothing changed but the resolution. Time spent iterating on Fast is not throwaway work.
Audio Included
Sound is generated with the picture at no additional cost, the same as on the other Seedance tiers. Your draft has a soundtrack, so you can judge pacing rather than guessing at it.
Failed Runs Are Refunded
Credits are held at submission and returned automatically when a generation fails or the provider never responds — which matters on the tier where you deliberately run the most attempts.
Where Fast Sits
Mini is for volume and rough shape at 5 credits a clip. Standard is for delivery at 1080p and 4K. Fast is the working tier in between — good enough to judge, cheap enough to repeat.
How to Use Seedance 2.0 Fast
This page is built around a single workflow: develop here, deliver on Standard. Following it costs a fraction of iterating directly at 1080p, and produces the same final clip.
Upload Your References
Up to five images at 30MB each. Upload order sets the numbering — the first file becomes @Image1. Use the same files you intend to use on the final render so nothing shifts between tiers.
Write and Cast the Prompt
Assign each reference a role, then describe the action: "@Image1 for the rider, @Image2 for the desert road. He rides toward camera as dust lifts behind him." The prompt field takes up to 20,000 characters.
Iterate at 480p or 720p
Four seconds at 480p is 15 credits. Change one element per run — a reference, a camera move, a duration — until the casting and the motion are right. This is the part the tier exists for.
Move the Winner to Standard
Copy the finished prompt to Seedance 2.0 Standard, upload the same references in the same order, and render at 1080p or 4K. Nothing else needs to change.
720p Is the Ceiling Here
Fast tops out at 720p by design. If you find yourself wishing this page had a 1080p option, that is the signal that the prompt is finished and it is time to move up a tier.
Seedance 2.0 Fast
FAQ
What the Fast tier gives up, what it keeps, how it compares to Standard and Mini, and what a clip costs.
Resolution, and nothing else you can set. Fast stops at 720p while Standard goes to 1080p and 4K. Reference images, the @Image syntax, the 20,000-character prompt, all six aspect ratios, durations from 4 to 15 seconds and generated audio are identical on both pages.
Both, mildly. It is a distinct model tier tuned for throughput, so expect slightly less refinement in fine motion and detail alongside the lower ceiling. What it does not change is how the model reads your prompt or casts your references — which is why a prompt validated here still behaves the same way on Standard.
From 15 to 85 credits. At 480p: 15 credits for 4 or 5 seconds, 30 for 10, 40 for 15. At 720p: 25 for 4 seconds, 30 for 5, 60 for 10 and 85 for 15. The exact cost for your settings appears above the generate button before you submit.
About 40% off at matching settings. Five seconds at 720p is 30 credits here against 50 on Standard; fifteen seconds at 720p is 85 against 140. Against the 4K tier the gap is far larger — 85 credits versus 705 for the same fifteen seconds.
When you are exploring rather than refining. Mini costs 5 credits for five seconds at 480p against 15 here, so it is the right tier for testing twenty different ideas. Once you know which idea you are pursuing and want to judge how it actually looks, Fast is where that judgement becomes reliable.
For most social video, yes — vertical clips watched on a phone rarely justify more. It is the wrong choice for anything shown on a large display, cut alongside 1080p footage, or likely to be cropped in later, since cropping throws away resolution you no longer have.
No, references are optional and the model works from text alone. But reference casting is the reason to choose Seedance over the cheaper text-driven models on this site — without it, Grok Imagine or Wan 2.6 will produce a clip for fewer credits.
Try Seedance 2.0 Fast
Fifteen credits gets you four seconds and a real look at how the model cast your references.