Wan 2.5 (preview) Partner Nodes in ComfyUI

1080P generation with sound.

We would like to share the Partner nodes for Wan 2.5 (preview) series in ComfyUI. This is a major leap forward in both image and video generation for Wan model family!

Highlights

• Audio-Visual Sync: High-fidelity voices, ASMR, effects, music; supports Chinese, English, and dialects.

• 10s Videos: Double length for fuller storytelling.

• Instruction Following: Better natural language, camera moves, structured prompts.

• Video Quality: More dynamic, stable, cinematic; up to 1080P 24fps.

• ID Preservation: Stronger consistency in image-to-video.

• Audio Conditioning: Use audio as input with prompts or keyframes.

Get Started

1. Update ComfyUI to the latest.

2. Search for “Wan Text to Video”, “Wan Image to Video” or Wan image nodes in Partner nodes.

Example Video Output with Sound

Note: This release introduces substantial changes to the Wan model framework. The current preview version is still under further refinement, and we look forward to sharing the community's feedback with the research team.

Please don’t hesitate to share the feedback. And let’s look ahead together to the official release. Enjoy creating!