---
title: "GGUF Quantization (Q4 / Q5 / Q8) — ComfyUI Glossary & Node Definition"
description: "Block-quantized weight format that shrinks 24 GB Flux and Hunyuan Video models to fit 8 GB and 12 GB GPUs. Learn how GGUF Quantization (Q4 / Q5 / Q8) works inside ComfyUI node workflows on Windows."
canonical: https://comfy.ink/words/gguf-quantization
updated: 2026-10-17
---

# GGUF Quantization (Q4 / Q5 / Q8)

Block-quantized weight format that shrinks 24 GB Flux and Hunyuan Video models to fit 8 GB and 12 GB GPUs.

## Facts (2026-10-17)

- **Term:** GGUF Quantization (Q4 / Q5 / Q8)
- **In Short:** Block-quantized weight format that shrinks 24 GB Flux and Hunyuan Video models to fit 8 GB and 12 GB GPUs.
- **Applies To:** ComfyUI SD 1.5, SDXL, Flux & Video

## Definition: GGUF Quantization (Q4 / Q5 / Q8)

Block-quantized weight format that shrinks 24 GB Flux and Hunyuan Video models to fit 8 GB and 12 GB GPUs.

Originally created for llama.cpp, GGUF weights loaded via ComfyUI-GGUF preserve near-FP16 visual quality while cutting VRAM consumption by 50–70%.

## How to tune this in your ComfyUI workflow

Every parameter in ComfyUI is exposed directly on its node rather than hidden behind global settings tabs. You can inspect or override this behavior per pass in [ComfyUI Studio](https://comfy.ink/studio).

New to node graphs? Grab the Windows installer at [/download](https://comfy.ink/download) and open a built-in template from the workflow menu.

## FAQ

### What does GGUF Quantization (Q4 / Q5 / Q8) do in ComfyUI?

Block-quantized weight format that shrinks 24 GB Flux and Hunyuan Video models to fit 8 GB and 12 GB GPUs. Originally created for llama.cpp, GGUF weights loaded via ComfyUI-GGUF preserve near-FP16 visual quality while cutting VRAM consumption by 50–70%.

### Where do I adjust this setting on the canvas?

Check the corresponding loader or [KSampler](https://comfy.ink/nodes/ksampler) widget in your active workflow graph.

## Related

- [ComfyUI Glossary](https://comfy.ink/words)
- [KSampler & KSampler (Advanced)](https://comfy.ink/nodes/ksampler)
- [ComfyUI Studio Workspace](https://comfy.ink/studio)
- [ComfyUI VRAM Requirements for SDXL, Flux & Video (6GB–24GB)](https://comfy.ink/guides/comfyui-vram-requirements-guide)
- [Latent Space & LATENT Pin](https://comfy.ink/words/latent-space)

## Download

Free desktop app: [Download Comfy](https://comfy.ink/download)

Markdown mirror: https://comfy.ink/words/gguf-quantization.md