TEXT TO 3D GUIDES

Qwen 3D Coder ComfyUI Setup

Updated 2026-11-17 · Steps assume a working ComfyUI install; details defer to the official model repos

On this page
  1. Before you start
  2. Step 1 — install or update ComfyUI
  3. Step 2 — download the weights
  4. Step 3 — load and generate
  5. Per-GPU notes
  6. FAQ
⚠️ ComfyUI node layouts change between releases. Treat this page as the general path, and confirm node names against your ComfyUI version and the model repo's README.

1. Before you start

2. Step 1 — install or update ComfyUI

If you already run ComfyUI for image models, update to a recent release before adding a 3D model — older builds may miss loaders used by newer text-to-3D checkpoints. From your ComfyUI folder:

git pull then pip install -r requirements.txt

If you are starting from zero, grab the official ComfyUI portable build for your OS from its GitHub releases page and run it once to confirm the UI loads at 127.0.0.1:8188.

3. Step 2 — download the weights

Download only from the official Hugging Face repos:

VariantHugging Face repoUse when
4B base (Three.js)ccore/Qwen3.5-4B-3D-CoderYou want interactive web scenes
4B GLBccore/Qwen3.5-4B-GLB-3D-CoderYou want GLB files for Blender/Unity
1.7B GLBccore account, 1.7B GLB CoderYou have a smaller GPU

Place the weights in your ComfyUI models directory following the repo README's layout (the exact subfolder depends on which loader nodes your ComfyUI version provides for this model class).

4. Step 3 — load and generate

  1. Open ComfyUI and start from an empty graph (or a community workflow shared for this model, if the repo links one).
  2. Add the model loader node, select the downloaded checkpoint, and set your prompt — descriptive scenes with explicit layout and lighting work best.
  3. Queue the prompt. Three.js variants emit scene code you can paste into an HTML file; GLB variants emit a file you can open in any 3D viewer.
  4. Validate the output in a viewer before importing it anywhere else — see the GLB export guide.

5. Per-GPU notes

Community reports suggest the 4B model is comfortable near 9 GB VRAM with standard precision, and workable below that with quantized weights at some quality cost. If you are on 8 GB, prefer the 1.7B GLB variant or a quantized build — and close other GPU-heavy apps first. Exact numbers vary by backend and driver; treat any figure you read (including here) as a starting point, not a promise.

6. FAQ

Do I need ComfyUI to run Qwen 3D Coder?

No — you can run the model directly with transformers-style loaders from the Hugging Face repo. ComfyUI is the popular choice because it gives you a node graph, previews and reusable workflows.

Where do I download the weights?

Only from the official Hugging Face repos (ccore account: Qwen3.5-4B-3D-Coder and its GLB/1.7B variants). Avoid re-uploads — weights from unofficial mirrors can be modified or outdated.