Seedance 2.0
ByteDanceVideo Generation
89.0
Performance
★ 4.5
Rating
520
Reviews
Video GenerationProprietary
About
ByteDance's unified multimodal audio-video generation model using a dual-branch diffusion Transformer for synchronized visual and audio output at up to 1080p.
Strengths
Dual-branch diffusion Transformer processes visual and audio simultaneously for perfect sync — instruments match finger movements, dance aligns with music beats. Innovative @ reference system for combining person photos, dance videos, and background music. Up to 1080p cinematic output. Strong camera control with tracking, orbit, and dynamic transitions. Multiple input modalities: text, image, audio, and video. Competitive pricing through Volcengine.
Specifications
- Context window
- —
- Parameters
- —
Available On
Volcengine APISeedance Webfal.aiReplicate
Features
text to videoimage to videovideo to videonative audiocamera controlreference systemlip syncdance generation
Performance Trend
Benchmark score trends over time for the top 5 benchmarks.
Loading history...
Benchmarks
Scores from various benchmark tests; higher is better.
| Test | Score | Percentile | Source |
|---|---|---|---|
| VBench | 84.5 | p95 | seed |
