Bernini v2: Semantic Planning Video Model by ByteDance
ComfyUI Wiki
Bernini-Diffusers-v2 is ByteDance's semantic-planning video model with a Wan2.2 dual DiT renderer, packaged for ComfyUI with FP8 and NVFP4 weights.
B
Bernini v2
Video GenerationDiTByteDanceComfyUIByteDance's Bernini-Diffusers-v2 — the second diffusers release of the Bernini framework. A Qwen2.5-VL semantic planner produces plan tokens that a Wan2.2 dual DiT renderer turns into video, supporting text-to-video, reference-to-video, video editing, and aesthetic transfer in a single workflow.
| Developer | ByteDance |
| Release Date | 2026-08-13 |
| Architecture | Qwen2.5-VL semantic planner + Wan2.2 dual DiT renderer (high/low noise, switch at 0.875) |
| License | Apache-2.0 |
| Variants | FP8 scaled · NVFP4 (both DiTs) via the ComfyUI pack |
| Capabilities | T2V · R2V · V2V editing · reference-guided editing · aesthetic transfer |
Guides and workflows related to this model series.
No articles found.
Comments
Sign in with GitHub to join the discussion.