ComfyUI-GGUF-Qwen3VL-TE
A ComfyUI add-on that patches ComfyUI-GGUF so Qwen-Image-2.1 models can run fully as GGUF files, fixing a text-encoder mismatch and missing architecture detection until upstream support arrives.
ComfyUI-GGUF-Qwen3VL-TE is a small add-on for ComfyUI-GGUF that lets Qwen-Image-2.1 GGUF models run without errors, until the upstream project supports them directly. It fixes two specific failures: a text-encoder mismatch that happens because ComfyUI-GGUF only loads the vision tower needed for an older Qwen model family, and missing architecture detection for Qwen-Image-2.1 GGUF files that lack their own metadata, such as those produced by certain third-party quantizers.
It adds no new nodes and changes no normal workflow; a user still loads the text encoder and the diffusion model with the existing ComfyUI-GGUF loaders, and the add-on patches the loading behaviour in the background, printing a confirmation to the console on startup. The vision tower it loads for the text encoder is used for real image-editing reference inputs, not only to satisfy the loader.