Seedance 2.0 Video Model
ByteDance’s video model on Labnana, in two tiers — Pro for the render you keep, Fast for the drafts that get you there.
- First and end frame control
- Up to 1080p on Pro
- 4–15 second clips
- Prompts up to 1,500 characters
Seedance 2.0 is ByteDance’s video generation model, available on Labnana in two tiers that take identical inputs. Pro is the fidelity tier and the only one that reaches 1080p; Fast trades that tier away for a shorter wait, which makes it the one to iterate on. The same prompt runs on either, so comparing them is a single dropdown change.
Both tiers accept a first frame, an end frame, or neither. Give them both frames and the model interpolates the shot between them, which is the closest thing here to directing a camera move rather than describing one.
Examples
What you can do with Seedance 2.0 on Labnana
01
Animate a still you already have
Upload an image as the first frame and the composition is settled before the model starts. Spend the prompt on what moves and how the camera travels, not on re-describing the scene.
02
Pin the ending with an end frame
Seedance is the model family here that accepts a closing frame. Give it a start and an end and it fills the shot in between — useful for transitions, reveals and before/after cuts.
03
Keep a subject recognisable
Attach up to nine reference images of one person, product or outfit and the model holds them across the clip instead of reinventing them each run.
04
Draft on Fast, render on Pro
Find the prompt on Fast at 480p or 720p, then switch the tier and re-run for the 1080p version. Nothing in the form is cleared when you switch.
Seedance 2.0 tiers at a glance
The generator filters itself to whatever the selected tier supports. Credit cost depends on tier, resolution and duration, and is shown live before you submit — so it is not duplicated here.
| Tier | Duration | Resolution | Aspect ratios | End frame | Prompt limit | Best for |
|---|---|---|---|---|---|---|
| Seedance 2.0 ProByteDance | 4–15s | 480p / 720p / 1080p | 6 (16:9, 4:3, 1:1, 3:4, 9:16, 21:9) | Supported | 1,500 characters | Final renders, faces and people, 1080p delivery |
| Seedance 2.0 FastByteDance | 4–15s | 480p / 720p | 6 (16:9, 4:3, 1:1, 3:4, 9:16, 21:9) | Supported | 1,500 characters | Drafts, prompt iteration, quick turnaround |
How it works
1
Pick the scene, not the settings
Start on image to video if you have a still, text to video if you only have a sentence, or reference to video if a specific subject has to survive the shot.
2
Choose the tier
Pro and Fast sit under Seedance 2.0 in the model dropdown. Resolution, duration and aspect ratio adjust to the tier you select — illegal values snap to the nearest legal one rather than failing at submit.
3
Change one parameter per run
Hold the prompt and move the tier, or hold the tier and rewrite one clause. Moving both leaves you unable to say which change did the work.
Where to use Seedance 2.0
All modelsAll three video scenes run on Seedance 2.0 Pro and Fast. They differ in what you bring to them.
Video generation on LabnanaAI Image to Video GeneratorYour image becomes the first frame — a short prompt sets the motion and the camera. Upload a still, pick a model, and get a watermark-free MP4 back.
Video generation on LabnanaAI Text to Video GeneratorWrite the shot you want — subject, setting, camera move — and the model builds every frame. Nothing to upload.
Video generation on LabnanaAI Reference to Video GeneratorShow the model who or what it is filming. Add reference images, describe the shot, and the subject stays recognisable from the first frame to the last.
HappyHorse 1.1 · AlibabaHappyHorse 1.1 Video ModelAlibaba’s video model on Labnana — the widest frame coverage here, and the only one that will render three seconds.
Frequently asked questions
Three ways to use Seedance 2.0
Upload a still, write a shot, or bring reference images — all three run on Pro and Fast.