FLUX.3 Image Klein renders 1024×1024 images in 2 seconds
Black Forest Labs has released FLUX.3 Image Klein, an open model that unites text-to-image, multi-reference editing and a new way to direct composition. Instead of hoping your prompt lands, you can now decide exactly where each element goes, and the way it works may surprise you.
FLUX.3 Image Klein renders 1024×1024 images in 2 seconds
Black Forest Labs' FLUX.3 Image Klein runs in Hugging Face Diffusers from day one. One model handles text-to-image and multi-reference image-to-image, and it adds a new capability: layout control. You write bounding boxes into the prompt to say where each part of the image should go, or which parts of a reference image to move. The distilled checkpoint makes a 1024 × 1024 image in about 2 seconds on an H100.
Both open checkpoints are in the FLUX.3 Image Klein collection on the Hub: black-forest-labs/flux-3-image-klein and black-forest-labs/flux-3-image-klein-base. You can also try the model in the Hugging Face Space at black-forest-labs/flux-3-image-klein to generate images without any setup.
Highlights
The release bundles several features:
One model for text-to-image, multi-reference image-to-image, and bounding-box layout control.
The distilled checkpoint makes a 1024 × 1024 image in about 2 seconds on an H100.
NVFP4 and FP8 weights load with from_single_file, for smaller GPUs.
LoRA training scripts for both text-to-image and image editing.
It ships as a Modular Diffusers pipeline, so you can pop blocks out or add your own.
Getting started
Install the latest diffusers and transformers (for the Qwen3-VL text encoder):
pip install -U diffusers transformers
Load the distilled checkpoint and generate an image:
import torch from diffusers import ModularPipeline pipe = ModularPipeline.from_pretrained("black-forest-labs/flux-3-image-klein") pipe.load_components(dtype=torch.bfloat16) pipe.to("cuda") pipe(prompt="a red fox", output="images")[0].save("fox.png")
What FLUX.3 can do
FLUX.3 responds best to a full scene description with the subject, the setting, the light, the composition, and the style. The official prompting guide from Black Forest Labs has more tips. The distilled checkpoint generates a 1024 × 1024 image in 16 steps by default. Pass height, width, or num_inference_steps to change them.
prompt = ( "A wide-angle wildlife photograph of a red fox standing alert in a snow-covered forest clearing at golden hour. " "The fox is centered in the frame, its thick orange winter coat contrasting with the fresh white snow, breath " "visible in the cold air. Snow-laden pine trees fill the softly blurred background, and low warm sunlight rakes " "across the scene from the left, casting long blue shadows." ) fox = pipe( prompt=prompt, generator=torch.Generator().manual_seed(0), output="images" )[0] fox.save("fox.png")
Image-to-image editing
Pass a reference image (or more) with the image argument. Write the prompt as a caption of the image you want, with the change included.
prompt = ( "The same red fox in the same snow-covered forest clearing, now at night under a full moon: cool blue " "moonlight, deep shadows, stars visible through the pine branches, the fox's eyes catching the light." ) night = pipe( prompt=prompt, image=fox, generator=torch.Generator().manual_seed(0), output="images", )[0] night.save("night.png")
Multiple references
With several references, refer to them in the prompt as `, , ... in the order they are passed to image`:
red_fox = pipe(prompt="a red fox", generator=torch.Generator().manual_seed(0), output="images")[0] bench = pipe(prompt="a wooden bench", generator=torch.Generator().manual_seed(0), output="images")[0] combined = pipe( prompt="The fox from sitting on the wooden bench from in a snowy setting.", image=[red_fox, bench], output="images", )[0] combined.save("combined.png")
Layout control with bounding boxes
Add a JSON list of regions after the caption. Each region has an id, a bbox, and a desc, and the caption mentions it as “.
Boxes are written as [y_min, x_min, y_max, x_max], y first, on a 0–1000 scale. Because the scale is normalized, the same box works at any image size.
layout_prompt = ( "Minimalist graphic illustration featuring a black silhouette of a person centered against a " "solid, vibrant chartreuse background . The figure is captured in a dynamic, mid-stride running " "pose, facing toward the left side of the frame. " '[{"id": "background_1", "bbox": [0, 0, 1000, 1000], "desc": "A flat field of neon yellow-green with a subtle ' 'paper texture and fine, organic grain."}, {"id": "silhouette_1", "bbox": [150, 150, 850, 850], "desc": "A ' "black silhouette of a person in motion, made of dense stippled ink. The leading edges are solid; the trailing " 'arm and leg dissolve into coarse square pixels and scattered dots."}]' ) runner = pipe(prompt=layout_prompt, generator=torch.Generator().manual_seed(0), output="images")[0]
Layout-based edits
For edits, each region says where its content comes from (src_bbox in the reference) and where it goes (tgt_bbox in the output). A region with the same two boxes stays where it is.
This example moves the knitted knight up and to the left, and pins the backdrop, the cliff, the needle, the dragon, and the fire:
from diffusers.utils import load_image knight_image = load_image("knight_orig.png") edit_prompt = ( "In , move the miniature grey amigurumi knight figure upwards and to the left along the " "yarn cliff . Keep the rest of the image completely unchanged, preserving the blurred yarn backdrop " ", the large knitted dragon , and the fiery orange thread . " '[{"id": "knight_1", "from": "ref_image_0", "src_bbox": [500, 150, 850, 350], "tgt_bbox": [194, 55, 544, 255], ' '"desc": "A miniature amigurumi knight crocheted from thick grey woolen yarn."}, ' '{"id": "backdrop_1", "from": "ref_image_0", "src_bbox": [0, 0, 1000, 1000], "tgt_bbox": [0, 0, 1000, 1000], ' '"desc": "A soft, out-of-focus backdrop of cream and indigo knitted yarn."}, ' '{"id": "cliff_1", "from": "ref_image_0", "src_bbox": [300, 0, 1000, 450], "tgt_bbox": [300, 0, 1000, 450], ' '"desc": "A steep cliff of stacked yarn skeins in mustard, terracotta, and indigo."}, ' '{"id": "needle_1", "from": "ref_image_0", "src_bbox": [420, 250, 550, 450], "tgt_bbox": [420, 250, 550, 450], ' '"desc": "An oversized polished steel tapestry needle."}, ' '{"id": "dragon_1", "from": "ref_image_0", "src_bbox": [100, 600, 950, 1000], "tgt_bbox": [100, 600, 950, 1000], ' '"desc": "A large, fluffy knitted dragon in terracotta and mustard yarn."}, ' '{"id": "fire_1", "from": "ref_image_0", "src_bbox": [350, 400, 650, 650], "tgt_bbox": [350, 400, 650, 650], ' '"desc": "Curled puffs of fiery orange angora thread suspended in mid-air."}]' ) moved = pipe(prompt=edit_prompt, image=knight_image, generator=torch.Generator().manual_seed(0), output="images")[0]
The source includes a figure comparing the reference image, with the knight's source box, against the output, with an arrow from the source box to the target box. The pinned regions are outlined in white. The rest of the scene stays put.
Base model and quantized variants
Everything above uses the distilled checkpoint. The base checkpoint runs 50 steps with classifier-free guidance, and it's the one to fine-tune (see the LoRA section below).
Black Forest Labs also ships the transformer as a single file, in bf16 and in two quantized formats: fp8r (rowwise FP8) and nvfp4. Load any of them with from_single_file, then load the other components by name so the bf16 transformer isn't downloaded too.
For the unquantized single file, use https://huggingface.co/{model_id}/blob/main/model.safetensors and drop quantization_config.
Which variant gets fast kernels depends on your GPU:
fp8r: Fast kernels on compute capability 8.9+ (Ada, Hopper, Blackwell). On other GPUs, weights are dequantized each step: lower memory, no speedup.
nvfp4: Fast kernels on Blackwell, with flashinfer-python and flashinfer-cubin at the same version. On other GPUs, the behavior is the same as above.
LoRA finetuning
FLUX.3 works for both text-to-image and image-to-image, which makes it a strong base for custom LoRAs. diffusers ships DreamBooth LoRA training scripts for both tasks. You can also train with Ostris' AI Toolkit.
--gradient_checkpointing recomputes activations in the backward pass (lower VRAM, slightly slower).
--cache_latents pre-encodes the images and frees the VAE.
--use_8bit_adam keeps the optimizer states in 8-bit.
A couple of knobs are worth experimenting with. The article includes a grid that trains the same tarot style with different LoRA ranks and optimizers. The adapters are on the Hub at diffusers-modular/flux3-tarot-lora. More training features and configurations are documented in the [DreamBooth FLUX.3 README](https://github.com/huggingface/diffusers/blob/main/examples/d
