Wan2.2-I2V-A14B-GGUF

Wan2.2-I2V-A14B-GGUF

Provides GGUF files of Wan2.2-I2V-A14B, a 14B model that turns a still image into a video, for use in ComfyUI with the ComfyUI-GGUF node. For people making short videos on a home GPU.

Made by QuantStack on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: This repository holds GGUF files converted straight from Wan-AI/Wan2.2-I2V-A14B, an image-to-video model with about 14.3B parameters. You give it a picture and a text prompt, and it makes a short video that starts from that image.

Quantized GGUF files are smaller than the original weights, so the model can run on less memory. At 4-bit, the weights are estimated at about 7.1 GB.

Who it suits: It suits creators and developers who already use ComfyUI and want to animate still images with an open video model on their own hardware.

Industry | Who uses it | How they use it | Example
Video | Creators | Animate a still picture into a short clip | Bring an illustration to life
Marketing | Designers | Make short motion from product images | Animate a product photo
Games | Artists | Draft motion for concept art | Test a character movement
Research | ML engineers | Try quantized video models | Compare quant sizes

If you do not use ComfyUI or have a GPU, you can skip it, since the card gives only the ComfyUI route.

Where to use it: The card names the ComfyUI custom node ComfyUI-GGUF, by city96, as the way to load these files. ComfyUI runs on Windows, macOS and Linux.

• ComfyUI with the ComfyUI-GGUF custom node
• The unet folder inside ComfyUI's models folder
• English and Chinese prompts, as the card lists both

Use case | Who | What they get back
Image to video | Creators | A short video from one starting image
Prompted motion | Designers | Movement guided by a text description
Quant testing | Engineers | A smaller model to run on limited memory

How to set it up: You need ComfyUI, the ComfyUI-GGUF custom node and Python with pip to fetch the files. The node's own readme covers its installation.

• ComfyUI installed
• The ComfyUI-GGUF custom node
• Python with pip, and disk space for the GGUF file you pick

1. Install the Hugging Face hub package.
2. Download the model files.
3. Place the model files in the unet folder under ComfyUI's models folder.
4. Install the ComfyUI-GGUF node by following its readme.

pip install -U huggingface_hub
hf download QuantStack/Wan2.2-I2V-A14B-GGUF

How to use it: In ComfyUI you load the GGUF file through the ComfyUI-GGUF node, add your starting image and a text prompt, and run the workflow. The video comes out of that graph. The card gives no run command beyond placing the files.

Example: You load a still photo of a street, write a short prompt that asks for people walking and a gentle camera move, and run the graph. You get back a short video that starts from your photo.

For newcomers and experts: If you are new, first get one picture and a one-line prompt working in ComfyUI. Then change the prompt and see how the motion changes.

If you know the field, the files are a direct GGUF conversion of the original Wan2.2 image-to-video weights. The GGUF node loads them in place of the standard diffusion model loader.

Place the GGUF file in the unet folder
Load it with the ComfyUI-GGUF node
Add an image and a prompt
Run the graph to get a video

Good to know: Because this is a quantized copy, all the original licensing terms and usage restrictions of the source model remain in force. Read those terms before you publish anything you make.

• License: apache-2.0, with the original model's terms still applying
• Cost to run: none named, but it needs a capable GPU
• Data it reaches: files you download, and the images and prompts you use
• Not for: use outside the source model's license terms

On Hugging Face, 13 November 2026

• Task: image to video
• Library: gguf
• Parameters: 14.3B
• Weights: about 7.1 GB at 4-bit (estimate)
• Formats: GGUF
• Downloads in the last 30 days: 1,140,540
• Likes: 464
• License: apache-2.0
• Languages: en, zh
• Based on: Wan-AI/Wan2.2-I2V-A14B
• Tags: gguf, image-to-video

Links

• Model: https://huggingface.co/QuantStack/Wan2.2-I2V-A14B-GGUF
• Files: https://huggingface.co/QuantStack/Wan2.2-I2V-A14B-GGUF/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download QuantStack/Wan2.2-I2V-A14B-GGUF

Then follow the usage in its model card. Nothing in this kit runs by itself.