Seedance 2.0
Released Feb 12, 2026
ByteDance Seed's joint audio-video generation model, sold as Dreamina Seedance 2.0: takes text, image, audio and video inputs for multimodal reference, editing and extension, at 4-15 seconds.
- Price per second
- —
- Max clip length
- 15 s
- Max resolution
- 4K
- Native audio
- Yes
- AA arena Elo
- 1136
- LMArena Elo
- 1479
- Capabilities
- text to video, image to video, video editing, extension
- Licence
- Proprietary
Strengths & weaknesses
Strengths
- Joint audio-video generation from text, image, audio and video inputs
- Video editing and extension, plus omni-reference (up to 9 images, 3 videos, 3 audio clips)
- 4K output (10-bit HEVC) on BytePlus ModelArk
- Seedance 2.0 Fast variant for low latency
Weaknesses
- 15-second maximum
- Text + audio and audio-only inputs not supported
- 4K is heavily rate-limited (15 RPM, concurrency 1)
ByteDance Seed's own paper page says Seedance 2.0 was 'officially released in China in early February 2026'. The exact date, 2026-02-12, comes from IT Home's report of the official ByteDance Seed WeChat announcement; Seed's English pages give no day. Resolution conflict: the Seed paper says native 480p/720p, but BytePlus ModelArk's model list offers 480p/720p/1080p (8-bit) and 4K (10-bit), so maxResolution is 4K per the API docs. Price is per token: ModelArk charges USD per million tokens (480p/720p $7.0 without video input, $4.3 with; 1080p $7.7 / $4.7; 4K $4.0 / $2.4 as listed), so perSecondUsd is null and nothing was converted. dreamina-seedance-2-0-fast-260128 is the Fast variant (480p/720p $5.6 / $3.3). No Volcano Engine (China) model ID verified. No open weights.
Links
API model strings
- byteplus-modelark
dreamina-seedance-2-0-260128 - byteplus-modelark
dreamina-seedance-2-0-fast-260128