Dreamina Seedance 2.0
Video modelPrevious generationCard summarizedOur next-generation video creation model
Source: Seedance 2.0: Advancing Video Generation for World Complexity (technical report) (26 pp)
Where to call it
- BytePlus ModelArk
dreamina-seedance-2-0-260128
Good for
- Complex multi-person motion and interaction that follows physics (sports, dance)
- Reference-driven generation from mixed images, video and audio
- Targeted video edits and extensions
Watch out for
Overview
ByteDance · 02Seedance, front
Seedance 2.0
Our next-generation video creation model
Typical price
$0.15 / sec
720p, 16:9, no video input
- Max resolution
- 4K
- Max length
- 15 s
- Modes
- text-to-video, image-to-video (first / first-last frame)
- Released
- 12 Feb 2026
Seedance 2.0 is a native multimodal audio-video generation model released in China in early February 2026. A unified architecture takes text, image, audio and video and adds reference-based generation, editing and extension. On ByteDance's SeedVideoBench 2.0 it rates first on all six text-to-video and image-to-video dimensions against Kling 3.0, Veo 3.1 and Sora 2 Pro, with the largest margins on audio, and it ranked #1 on Arena.AI's text- and image-to-video leaderboards (8 Apr 2026). §1, §2.2–2.4
dreamina-seedance-2-0-260128- Typical price
- $0.15 / sec
- Max resolution
- 4K
- Max length
- 15 s
- Aspect ratios
- 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
- Frame rate
- 24
- Released
- 12 Feb 2026
- text-to-video
- image-to-video (first / first-last frame)
- omni reference-to-video (up to 9 images, 3 videos, 3 audio clips)
- video editing
- video extension (forward/backward, up to 3 clips stitched)