Flux 3 Klein GGUF

FLUX 3 Image Klein is a 5.6B parameter open-weights diffusion transformer from Black Forest Labs for text-to-image generation and reference-based image editing. A single model accepts a text prompt and up to 10 reference images to produce a new image or edit existing inputs. It also supports bounding-box coordinates in the prompt for spatial layout control and text rendering. This repository provides GGUF quantizations of both the 16-step Distilled (sd) and 50-step Base (base) checkpoints for ComfyUI.

• Loader nodes: GitHub / ComfyUI Registry
• Further documentation: ComfyUI-GGUF-Reboot Maintainer's docs

Example

Prompt used for the image above:

Full-color outdoor photograph at night, shot from a low Dutch angle with the horizon tilted sharply across the frame. A retro hotel entrance glows in teal and orange beneath a deep navy sky, its facade washed by cyan neon tubing and warm amber light spilling through broad glass doors. Across the upper center, large gold block letters mounted above the doorway read "Flux3 Klein GGUF", each letter edged by a narrow cyan glow. In the foreground, a Black woman bellhop in her thirties with dark skin and black hair pinned into a neat bun wears a vintage burgundy uniform jacket with brass buttons, a white shirt, striped tie, peaked cap, and matching trousers. She leans forward while pulling a polished gold luggage cart diagonally toward the right, both hands gripping its curved handle as it rolls across rain-darkened pavement. The cart carries three worn suitcases: a tall brown leather case with brass corners on top, a scuffed tan suitcase below, and a smaller oxblood case tied with cord near the front. Water beads cling to the metalwork, and long cyan and orange reflections stretch across puddles around the wheels. Inside the brightly lit lobby, teal-toned glass walls reveal two women in vintage uniforms standing beside a wooden reception desk. On the left is an East Asian woman in her twenties with fair skin and short black hair, wearing a cream blouse and fitted teal vest; beside her stands a white woman in her forties with fair skin and curled auburn hair, wearing a pale blue dress and matching waistcoat. Both face inward behind the desk. Potted palms and broad tropical leaves rise between them, while a seated figure in a dark coat rests in an armchair beyond the counter. Warm pendant lamps illuminate the desk's brass lamp, ledger, and telephone. Cyan window frames cast hard diagonal highlights over the wet sidewalk, with scattered droplets catching points of neon light along the lower edge.

Prompting tips for this model:

• Write full scene descriptions instead of short tag lists, or run the Qwen3.5 9B prompt upsampler node in ComfyUI to expand short instructions automatically.
• For spatial layout control or regional edits, include [top, left, bottom, right] coordinates on a 0 to 1000 grid directly in the prompt, for example place a brass lantern at [620, 180, 860, 340].

Available Quantizations

Q8_CR is recommended on GPUs with native INT8 acceleration. For lower VRAM usage under 3 GB, Q4_CR_L uses packed INT4 ConvRot weights with INT8 activations and keeps additional sensitive attention projections in Q8_CR, while Q4_CR uses INT4 weights and INT4 activations. One-dimensional normalization tensors and final_layer.image.linear stay in FP32, and the 128-wide input patch projections (emb_in) remain in source BF16.

Distilled (flux3_image_klein_sd)

Base (flux3_image_klein_base)

Model Input and Output

Variants and inputs

Outputs

ComfyUI Setup

Load the model with the UnetLoaderGGUF node from comfyui-gguf-reboot. The loader node gives the best performance. Set CLIPLoader to the flux3 type when loading the Qwen3-VL 4B text encoder.

Dependencies

Place the GGUF file and these companion files from the upstream repository in the ComfyUI model folders before running a workflow.

Workflow downloads

Import these JSON files into ComfyUI:

• Text-to-Image workflow
• Image Edit workflow