Base model / SOURCE CHECKED 12 OCT 2026
Qwen-Image 2.1
Before you download.
This Comfy-Org repository repackages Qwen-Image 2.1 for ComfyUI. Start with its linked workflow and matching diffusion model, text encoder and VAE.
Local setup checklist
- Read the upstream licence before choosing a commercial use case. Download from the linked publisher rather than an unknown mirror.
- Use the repository’s text-to-image or image-edit workflow. Match each loader to the downloaded model version.
- Put diffusion weights in models/diffusion_models, the encoder in models/text_encoders and the VAE in models/vae.
- Run a small, non-sensitive test. Record workflow version, GPU, peak memory and generation time before increasing resolution.
Memory and limitations
No minimum VRAM independently measured here. Total memory includes the encoder, VAE and intermediate tensors.
Licence before use
Qwen Research — review full terms. Check the full current licence; labels alone do not establish commercial permission.
Frequently asked questions
Is this an uncensored model?
This listing makes no uncensored claim. Behaviour and restrictions depend on the model, workflow and upstream terms.
Does WildWeights run the model?
No. WildWeights provides a directory and setup guidance. Generation happens in your own environment.
Primary source
Publisher model card and files ↗
Guidance is based on the source review, not our own inference benchmark. No user files or prompts are sent to a model by this site.
← Explore other workflows