MOSS-Image Base & SFT (INT6 W6A8 & INT4 W4A8) for ComfyUI
Repackaged model files for ComfyUI with low-bit quantized versions of both MOSS-Image-Base and MOSS-Image-SFT.
The original models are:
The official ComfyUI repackage (BF16 and INT8 ConvRot diffusion models, text encoders and VAE) is at Comfy-Org/MOSS-Image.
Files
diffusion_models/moss_image_base_W6A8.safetensors: the MOSS-Image-Base diffusion transformer as a single file in ComfyUI's native INT6 W6A8 format.diffusion_models/moss_image_base_W4A8.safetensors: the MOSS-Image-Base diffusion transformer as a single file in ComfyUI's native INT4 W4A8 (codebook) format.diffusion_models/moss_image_sft_W6A8.safetensors: the MOSS-Image-SFT diffusion transformer as a single file in ComfyUI's native INT6 W6A8 format.diffusion_models/moss_image_sft_W4A8.safetensors: the MOSS-Image-SFT diffusion transformer as a single file in ComfyUI's native INT4 W4A8 (codebook) format.
The text encoder and VAE are not included here. Download them from Comfy-Org/MOSS-Image:
text_encoders/qwen3.5_9b_int8_convrot.safetensorsortext_encoders/qwen3.5_9b_w4a8_gptq.safetensorsvae/flux2-vae.safetensors
Model storage paths
Place the files in the following folders:
📂 ComfyUI/
├── 📂 models/
│ ├── 📂 diffusion_models/
│ │ ├── moss_image_base_W6A8.safetensors
│ │ ├── moss_image_base_W4A8.safetensors
│ │ ├── moss_image_sft_W6A8.safetensors
│ │ └── moss_image_sft_W4A8.safetensors
│ ├── 📂 text_encoders/
│ │ ├── qwen3.5_9b_int8_convrot.safetensors
│ │ └── qwen3.5_9b_w4a8_gptq.safetensors
│ └── 📂 vae/
│ └── flux2-vae.safetensors
Load the diffusion model with ComfyUI's Load Diffusion Model node, the same way as the Comfy-Org files. The tensor names match Comfy-Org's repackage, and the quantization settings are stored in ComfyUI's native _quantization_metadata header, so no extra loader is needed.
The original model recommends 50 steps, CFG 3, shift 5, and negative-branch skipping at layer 22. I've have been finding good results at a minimum of 25 steps, CFG 3, and negative-branch skipping at layer 22 (with the shift node at 5 the results are identical without the node)
Conversion details
The large layers of all 42 transformer blocks (168 layers) are stored at 6 bits (W6A8) or 4 bits (W4A8) instead of 16 bits. Small, sensitive parts such as the embedders, modulation, output layer and norms stay at full precision. Converted with my ComfyUI-INT8-Fast (fork). Requires comfy-kitchen >= 0.2.37 on ComfyUI v0.38.0 or newer.
How it compares
| Format | File size | Size vs BF16 | Size vs INT8 | Weight error vs BF16 |
|---|---|---|---|---|
| BF16 (Comfy-Org) | 10.2 GB | 100 % | 188 % | 0 % |
| INT8 ConvRot (Comfy-Org) | 5.4 GB | 53 % | 100 % | ~1 % (estimated) |
| W6A8 (this repo) | ~4.4 GB | ~43 % | ~81 % | 2.3 % |
| W4A8 (this repo) | ~3.3 GB | ~33 % | ~62 % | 7.3 % |
Image prediction difference vs BF16 (SFT model):
| Format | First step (pure noise) | By step 4 |
|---|---|---|
| W6A8 | 5.6 % | 3.2 % |
| W4A8 | 13.2 % | 9.5 % |
- W6A8: about 19 % smaller than INT8 and very close to BF16. Images should be nearly identical.
- W4A8: about 38 % smaller than INT8, with roughly 3× the error of W6A8. Expect more visible differences in layout and fine detail.
- The difference is largest at the first, noisiest step and shrinks after that.
Check output quality for your own prompts and workflows before relying on it.
License
The original model is released under the Apache License 2.0.