Skip to content
Follow me on X
•ComfyUI Wiki
NewsSVI 2.0 Pro Released - Infinite-Length Video Generation with Wan 2.2 Support

SVI 2.0 Pro Released - Infinite-Length Video Generation with Wan 2.2 Support

12/29/2025

SVI 2.0 Pro Released - Infinite-Length Video Generation with Wan 2.2 Support

Stable Video Infinity Logo

Stable Video Infinity (SVI) 2.0 Pro has been officially released, adding support for the Wan 2.2 base model. Users can now use the new LoRA models in ComfyUI to generate infinite-length video content.

Major Updates

Wan 2.2 Base Model Support

SVI 2.0 Pro is built on the Wan 2.2 base model, providing better video generation quality and more stable performance. This version further optimizes video generation consistency and scene transition effects while maintaining the original functionality of SVI 2.0.

Multiple LoRA Model Options

The new version offers multiple LoRA model options, allowing users to choose the appropriate version based on their needs:

  • HIGH Version: Provides higher video quality and detail performance
  • LOW Version: Offers faster generation speed while maintaining good quality

These models are all based on the Wan 2.2 I2V A14B architecture and support generating video content from a single image.

Community Showcase

Many users have already started creating with SVI 2.0 Pro and sharing their work. Below are some video examples generated by community users using SVI 2.0 Pro workflows:

cr: @PT

cr: @PT

cr: @Aiwood

cr: @高姿态的浅唱

cr: @wallen

cr: @CUDA out of memory

Technical Features

SVI 2.0 Pro inherits the core advantages of the SVI series, capable of generating video content of any length while maintaining high temporal consistency and smooth scene transitions. This technology is particularly suitable for scenarios requiring long video content generation, such as short video creation and storytelling.

Infinite Length Generation

Unlike traditional video generation methods, SVI 2.0 Pro has no video length limitations. Users can start from a single image and continuously generate video content, theoretically generating videos of any length.

High-Quality Scene Transitions

While maintaining single-scene stability, the new version also supports natural transitions between scenes, making the generated video content richer and more coherent.

ComfyUI Workflow Support

SVI 2.0 Pro fully supports ComfyUI workflows, allowing users to use these new models in a familiar interface. The community has created various workflow configurations, and users can choose the appropriate configuration based on their needs.

Model Access

Users can obtain SVI 2.0 Pro LoRA models through Hugging Face. The models provide multiple versions for users to choose from, including different configurations based on Wan 2.1 and Wan 2.2.

Future Plans

The development team continues to optimize SVI technology. The next planned release is Wan 2.2 Animate SVI, which will support skeleton-based animation video generation.

Related Links