GitHub radar
Viggle's fast add-on cuts Qwen-Image-2.1 to 6 steps
Viggle trained a speed add-on for Alibaba's Qwen-Image-2.1 that builds and edits images in 6 steps instead of 40, staying close to the original in quality.
This is a speed-focused add-on for Qwen-Image-2.1, the image model released by Alibaba's Qwen team. It was trained by Viggle, a company known for AI character and video tools, using a distillation method that teaches a shorter path to copy the output of the full model. The base model needs 40 steps to build an image; this version needs six. It handles both jobs the base model does: generating an image from a text description, and editing an existing photo with 1 to 3 reference images, including keeping a specific face or product recognizable. According to Viggle, quality stays close to the base model in most cases. The two places where it falls behind: small, dense text rendered inside an image, and complicated edits like swapping a face across several reference photos. The add-on loads on top of the original Qwen-Image-2.1 weights, so you still need the base model installed alongside it.
Why a vibe-coder should care
If you already generate images locally and wait through the full 40-step pass every time, this cuts most of that wait without giving up much quality for ordinary generation and editing. It will not help with fine printed text inside an image or with tricky multi-reference face swaps — the slower base model still does better there. The license is for non-commercial use only, so treat the results as tests and personal projects, not something to sell.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Set up the fast Viggle version of Qwen Image 2.1: https://huggingface.co/Viggle/Qwen-Image-2.1-viggle-turbo — load it on top of the base Qwen-Image-2.1 model via ComfyUI or diffusers, and show me how to generate an image in six steps instead of forty.
A regular laptop with 16 GB of RAM or more - same as the base Qwen-Image-2.1, about 4 GB in the quantized version. The add-on itself is small and barely adds to that.
Open on Hugging Face▌ More finds