GitHub radar

FLUX 3 Image Klein: draw and edit images from text prompts

FLUX 3 Image Klein is an image generation and editing model from Black Forest Labs (the creators of Flux). Comfy-Org packaged a ready-to-use build for ComfyUI — an image-editing tool — so you don't need to hunt down files yourself.

01black-forest-labs/flux-3-image-klein-comfy 865393k downloads/mo5.6B paramsimage-to-image

It's a 5.6-billion-parameter neural network: it draws a picture from your text description and can use up to 10 of your photos as a reference for style or blending. Comfy-Org packaged and pre-configured the build for ComfyUI (a visual tool for working with images), so you don't need to find the files separately.

When it helps

Use it if you want to generate and edit images locally on your own computer instead of an online service. Skip it if you don't have a graphics card with at least 8GB of memory.

Pros

  • The lightweight version (5.8GB) runs on a gaming GPU with 8GB memory.
  • A fast mode draws a picture in just 16 model steps.
  • It can place objects on a prompt grid and draw readable text.

Cons

  • Free for non-commercial use or companies earning under $5M a year.
  • Won't run without a GPU that has at least 8GB of memory.

How to set it up — step by step

  1. 1Open your AI agent (Claude Code, Codex, etc.) in the ComfyUI folder, or wherever you want to set it up.
  2. 2Send the agent the text from the block below.
  3. 3Tell the agent your GPU's memory size — for example, that it has 8GB.
  4. 4Wait while the agent installs the model and picks the version that fits your GPU.
  5. 5Check the result — ask the agent to generate a test image from a short text prompt.
  6. 6Try the fast 16-step mode to see how much quicker generation gets.

Text for your agent

Copy this and send it to your agent — Claude Code, Codex, any of them:

Set up the FLUX 3 Image Klein model packaged by Comfy-Org for ComfyUI: https://huggingface.co/black-forest-labs/flux-3-image-klein-comfy — pick the lightweight version for my GPU and show me how to generate a first image from text.

You need a GPU with 8 GB of VRAM — an ordinary gaming card like the RTX 3070 is enough.

Open on Hugging Face