Seedance 2.0 — the complete guide
Dreamina Seedance 2.0 is ByteDance's flagship AI video generation model — up to 4K resolution, reference-to-video generation, and the default engine behind AI Cineworks. Here's what it can do and how to use it.
Core specs
| Spec | Value |
|---|---|
| Max resolution | 4K (also 480p, 720p, 1080p) |
| Max clip length | 15 seconds |
| Generation modes | Text-to-video, image-to-video, reference-to-video |
| Reference assets | Up to 15 (9 images + 3 videos + 3 audio) |
| Aspect ratios | 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive |
| Variants | Standard, Fast (lower cost/latency), Mini (lowest cost) |
What it's good for
Text-to-video
Describe a shot in plain language — camera movement, lighting, subject action — and Seedance 2.0 generates a full clip from scratch. Best for concepting, previsualization, and generating footage you don't have a source image or video for.
Image-to-video
Animate a still image (a keyframe, a generated image, a photo) into motion, with a prompt guiding how the scene should move and evolve from that starting frame.
Reference-to-video
Feed in reference images, a reference video, and/or reference audio to anchor the output to a specific character design, location, camera motion, or soundtrack — useful for keeping consistency across multiple shots in the same project.
Pricing
Seedance 2.0 is billed by tokens, with rates that vary by resolution and by whether a video is included as input:
| Resolution | With video input | Without video input |
|---|---|---|
| 480p / 720p | $0.0043 / 1K tokens | $0.0070 / 1K tokens |
| 1080p | $0.0047 / 1K tokens | $0.0077 / 1K tokens |
| 4K | $0.0024 / 1K tokens | $0.0040 / 1K tokens |
Token count follows width × height × frame rate × duration ÷ 1024, so cost scales with resolution and clip length. Fast and Mini variants are priced lower for the same resolution tier.
Seedance 2.0 vs Seedance 2.5
Seedance 2.0 is still the model to reach for when you need higher resolution — up to 4K, where 2.5 tops out at 720p. Seedance 2.5 pulls ahead on clip length (30s vs 15s), reference capacity (50 vs 15 assets), and native video editing/extension. Read the full Seedance 2.5 guide for a side-by-side comparison.
Generate with Seedance 2.0 right now
Seedance 2.0 is the default model in AI Cineworks' Text-to-Video, Image-to-Video, and OnSet tabs — no API setup required.
Frequently asked questions
480p, 720p, 1080p, and 4K, making it the highest-resolution model in the Seedance family.
Up to 15 seconds per generation.
Supplying reference images, a reference video, or reference audio alongside a text prompt so the output matches a specific character, style, location, or motion — instead of generating purely from text.
Yes — Seedance 2.0 Fast and Seedance 2.0 Mini variants trade some quality for lower cost and faster generation, both available in AI Cineworks.
Yes — Seedance 2.0 is the default video model across AI Cineworks' Text-to-Video, Image-to-Video, and OnSet tabs.