Seedance 2.0
ByteDance
ByteDance's video generation model (2026), powering the Dreamina/Doubao platforms. Seedance 2.0 accepts quad-modal inputs (text, image, video, audio) and features a 'Universal Reference' system that locks composition, camera movement, and character actions across shots for consistent multi-shot sequences.
Strengths
- Quad-modal inputs: text, image, video, and audio
- Universal Reference locks composition and character actions across shots
- Strong multi-shot consistency
- Native audio in generated video
- Backed by ByteDance's serving scale
Caveats
- Proprietary - API/app access only
- Served mainly via third-party API partners
- Newer ecosystem than Western incumbents
- Vendor
- ByteDance
- License
- Proprietary
- Modalities
- videoaudio
Capabilities
Resources
Pricing
ByteDance publishes its own price on BytePlus ModelArk: dreamina-seedance-2-0-260128, online inference in USD per million tokens — 480p/720p 7.0 without video input, 4.3 with video input; 1080p 7.7 / 4.7; 4K 4.0 / 2.4; offline inference "Not supported yet". Worked per-video examples run about USD 0.39-9.33 per video depending on resolution and input length. Cheaper variants: dreamina-seedance-2-0-fast-260128 (5.6 / 3.3, time-limited 25% off) and dreamina-seedance-2-0-mini-260615 (3.5 / 2.1, time-limited 60% off, ~USD 0.03/sec at 720p during the promo).