Seedance 2.0 AI Video Generator
Seedance 2.0 is a video generation model from ByteDance that makes 4 to 15 second clips with optional sound
What is Seedance 2.0?
Seedance 2.0 is a video generation model from ByteDance that makes 4 to 15 second clips with optional sound: give it a text prompt, optionally with a first frame or up to 9 reference images, and the sound is generated alongside the picture if you want it. On blubiu it runs in three versions: Mini for cheap drafts, Fast, and the full 2.0 for 1080p.
Developer: ByteDance · Checked 2026-11-02
What Seedance 2.0 is good at
- +Keeping a character consistent: pass up to 9 reference images (plus up to 3 reference videos and 3 audio clips) and describe the character with the same words every time.
- +Clips that need sound: ambient noise, footsteps or dialogue are generated with the video, so a silent clip doesn't need a separate audio pass.
- +Drafting cheaply, then upgrading: try the shot with Mini at 480p, and rerun the version you like with 2.0 at 720p or 1080p.
- +Animating a still you already have: one image becomes the first frame in image to video, and a second image becomes the last frame for a first and last frame video.
- +Short story sequences: plan a scene as several clips that share the same reference images and the same character sentence, so every cut shows the same character.
When not to use it
- −You need one continuous shot longer than 15 seconds. Split the scene into clips (the Agent plans the cuts for you).
- −You need 2K output. Seedance tops out at 1080p; MiniMax H3 renders 2K. See MiniMax H3 →
- −You only need a fixed 8-second shot with sound and want first-and-last-frame control. Veo 3.1 Fast is built around exactly that. See Veo 3.1 Fast →
- −You already like a clip and only want it sharper. Running the prompt again gives a different take; upscale the one you have with FlashVSR instead. See FlashVSR →
Seedance 2.0 versions and specs
| Version | Inputs | Length | Resolution | Sound | Price |
|---|---|---|---|---|---|
| Seedance 2.0 Mini | Text, First frame (+ optional last), References (images ≤9, videos ≤3, audio ≤3) | 4–15 s | 480P / 720P | Optional | 480P 8.84 · 720P 19 credits/s, 4 s minimum; subscribers 40% off: 480P 5.31 · 720P 11.4 credits/s; reference video 480P 5.38 · 720P 11.58 credits/s |
| Seedance 2.0 Fast | Text, First frame (+ optional last), References (images ≤9, videos ≤3, audio ≤3) | 4–15 s | 480P / 720P | Optional | 480P 14.22 · 720P 30.56 credits/s, 4 s minimum; subscribers 40% off: 480P 8.54 · 720P 18.34 credits/s; reference video 480P 8.46 · 720P 18.18 credits/s |
| Seedance 2.0 | Text, First frame (+ optional last), References (images ≤9, videos ≤3, audio ≤3) | 4–15 s | 480P / 720P / 1080P | Optional | 480P 18.36 · 720P 39.46 · 1080P 98.42 credits/s, 4 s minimum; subscribers 40% off: 480P 11.02 · 720P 23.68 · 1080P 59.06 credits/s; reference video 480P 11.18 · 720P 24.02 · 1080P 59.84 credits/s |
Standard prices, read live from the platform catalog; once signed in, the Generate button shows your account's actual price. Failed generations are refunded.
How to prompt Seedance 2.0
One action per clip
Describe a single continuous action and when it happens ("0–2 s: steps forward, 2–4 s: looks up"). Several unrelated actions in 4 seconds come out rushed or skipped.
Lock the look in words
Write the character's appearance as one fixed sentence (shape, colors, distinctive features) and paste it into every clip. Reference images help most when the text agrees with them.
Use a first frame for composition
If the framing matters, generate or upload the opening image first and pass it as the first frame; the prompt then only needs to describe motion.
Say what you want to hear
Sound is generated too, so name it: "ambient forest sound, no music, no dialogue". Otherwise the model may add music you didn't ask for.
Let your uploads pick the mode
In the generator, one image is read as the first frame, two as the first and last frame, and three or more as references; adding a video or an audio clip switches to omni reference. To use one or two images as references instead, click the tag on the image.
Name each reference in the prompt
With several references, start by saying what each one is ("image 1 is the character, image 2 is the kitchen"), then describe the shot. Trim reference videos to the part you need: their seconds are billed too.
Seedance 2.0 examples
All generated on blubiu, unedited.
Seedance 2.0 FAQ
What is Seedance 2.0?
Seedance 2.0 is ByteDance's text- and image-to-video model. It generates 4–15 second clips from a prompt, a first frame or several reference images, and can generate matching sound. blubiu lets you use it directly with credits — this is not ByteDance's own app.
What's the difference between Mini, Fast and 2.0?
They share the same inputs and duration range. Mini is the cheapest per second and is meant for drafts; Fast costs a little more; the full 2.0 is the only one that renders 1080p. The table above lists each version's resolutions and current price.
How long can a Seedance 2.0 clip be?
Between 4 and 15 seconds per generation. For a longer video, generate several clips and join them; describe the end state of one clip as the start of the next so the cut feels continuous.
How is it priced on blubiu?
Per second of video at the resolution you pick, with a 4-second minimum; when you pass reference videos, their seconds are billed as well. The exact credits per second for each version are in the table above and on the Generate button before you confirm. Failed generations are refunded automatically.
Can I try Seedance 2.0 for free?
Yes, with the Free plan's monthly credits. Mini at 480p is the cheapest way to test a shot, and the Generate button shows the exact cost before you confirm. Failed runs are refunded automatically; plans and credit packs are on the pricing page.
Does it support a first and last frame?
Yes. Add one image and it opens the clip; add a second and the clip ends on it, so the model only invents the motion in between. Use first and last frame video for transitions, and keep the subject, light and camera height the same in both frames.
How do I keep the same character across clips?
Reuse the same reference images and the same description sentence in every clip. A consistent character sheet or a character turnaround sheet shows the character from several angles in one image; pass it on reference to video together with any pose or scene references.
Seedance 2.0 or Wan 3.0?
Both make clips with sound up to 15 seconds at 1080p. Pick Seedance when you want to draft on Mini and rerun the keeper on the full model. Pick Wan 3.0 for clips as short as 2 seconds, for replacing the person in a video, or when you are waiting on the result and want its faster Prime tier.
Can I upscale a clip to 2K or 4K?
Yes, after it is made. FlashVSR raises a finished clip to 1080P, 2K or 4K and keeps its motion, length and frame rate, so a Mini or Fast draft you like can be delivered without regenerating it. Run it on the AI video upscaler page.
When should I not use Seedance 2.0?
When you need 2K (use MiniMax H3), a single shot longer than 15 seconds (split it into clips), or the lowest possible price per second for rough drafts (Veo 3.1 Fast and MiniMax H3 Lite cost less per second).
Tools that use Seedance 2.0
Create Video with AI →
Create video with AI from a sentence, a first frame, first and last frames or up to 9 reference images: 1–15 second clips with Seedance, Veo, MiniMax and Wan.
AI Text to Video Generator →
Turn a written description into a 1–15 second video with Seedance, MiniMax H3, Veo and more — no images needed.
AI Image to Video Generator →
Upload an image as the first frame, describe the motion and generate a video with Seedance, Veo 3.1, MiniMax H3 Lite or Wan 3.0.
Pair Seedance 2.0 with
GPT Image 2 & 2.5 Image Generator →
Draw the opening still first, with the character and composition you want, then pass it here as the first frame and describe only the motion.
Consistent Character Sheet Generator →
Make one sheet with three face angles and a full turnaround, and reuse it as the reference image in every clip of the series.
AI Outfit Changer →
Put the same character in a new outfit and pose, then animate that image as the first frame of a clip.
AI Video Upscaler →
Draft on Mini or Fast, keep the take you like, and lift it to 1080P, 2K or 4K without generating it again.
Other video models
Use MiniMax H3, the Hailuo video model, for 2K clips with reference images and videos, or H3 Lite for the cheapest silent drafts from 1 second.
Generate an 8-second 720p clip with optional sound using Google's Veo 3.1 Fast, from text or from a first and last frame. One fixed price per clip.
Use Alibaba's Wan 3.0 for 2–15 second videos with sound from text, a first frame or reference images, or to replace the person in a video.
Run Kling Motion Control 2.6 online: a full-body photo copies the moves from a 3–30 second video, at 720P or 1080P. No Kling account needed; try it free.
Upscale an MP4 or MOV up to 120 seconds to 720P, 1080P, 2K or 4K with FlashVSR, a video super-resolution model. Billed per source second.
See image models →