FLUX.2 Klein · 4B
FLUX.2 klein 4B is a 4 billion parameter rectified flow transformer model from black-forest-labs, licensed under Apache 2.0. It is an image-to-image pipeline and also supports text-to-image and multi-reference editing. The model card describes it as the fastest image model in the FLUX.2 family, unifying generation and editing in a single compact architecture, with end-to-end inference in as low as under a second. It is built for interactive workflows, production deployments, and latency-critical applications, and can run on consumer GPUs with as little as 13GB VRAM. It supports English language input and is available via BFL API, ComfyUI, and Diffusers. The official model card notes limitations around factual information, text rendering accuracy, bias, prompt following, and certain out-of-scope uses.
What do you want to do?
Choose a task. Get the software, files, settings, and steps to run it.
Create images locally with FLUX.2 Klein
Generate images from text prompts in ComfyUI with the fast distilled or more adjustable base 4B model. Save your results and reuse the visual workflow.
Edit images locally with FLUX.2 Klein
Upload an image, describe the changes you want and save the edited result in ComfyUI. This reference-guided workflow supports a range of visual edits; it does not use an inpainting mask.
Open a guide for its software, hardware requirements, and setup steps.
At a glance
- Parameters
- 4B
- Architecture
- unknown
- License
- Apache 2.0
- Software
- ComfyUI, Diffusers
Will it run on your machine?
Save your machine to see a personalized rating and its reasoning.
Add my hardwareBest for
- Text-to-image generation
- Image editing
- Multi-reference editing
- Interactive workflows
- Latency-critical applications
Tradeoffs
- Not intended or able to provide factual information
- Text rendered in outputs may be inaccurate or distorted
- May represent or amplify biases from training data
- May fail to generate output matching prompts
- Prompt following is heavily influenced by prompting style
Ways to run it
Diffusers · See official guide
Repository-specific command in publisher documentation; confirm dependencies and hardware in the source.
import torch
from diffusers import Flux2KleinPipeline
device = "cuda"
dtype = torch.bfloat16
pipe = Flux2KleinPipeline.from_pretrained("black-forest-labs/FLUX.2-klein-4B", torch_dtype=dtype)
pipe.enable_model_cpu_offload() # save some VRAM by offloading the model to CPU
prompt = "A cat holding a sign that says hello world"
image = pipe(
prompt=prompt,
height=1024,
width=1024,
guidance_scale=1.0,
num_inference_steps=4,
generator=torch.Generator(device=device).manual_seed(0)
).images[0]
image.save("flux-klein.png")Explore its uses
Keep exploring
Qwen-Image 2.1
Generate and edit images, including transparent images and edits guided by reference images.
FLUX.1 schnell
A 12-billion-parameter image model designed for generation in one to four steps.
Cosmos3 Super · Text2Image
A large open-weight image generator aimed at physical-AI scenes, with structured prompt support and publisher-tested GB200 or multi-H100 deployment.