Flux.1 - AI Image Models Tool

Overview

FLUX.1 is the official inference codebase and model suite from Black Forest Labs for text-to-image generation and in‑place image editing. According to the project's GitHub repository, the repo provides minimal, production-oriented inference tooling, examples, and optional TensorRT support for running FLUX open-weight models locally or in server environments (installation and TensorRT instructions are in the README). The FLUX.1 family includes multiple released variants — notably FLUX.1 [schnell] (fast, Apache-2.0), FLUX.1 [dev] (guidance-distilled, gated dev license), and FLUX.1 Kontext [dev] (in‑context image editing) — plus a suite of editing helpers (Fill, Canny, Depth, Redux) available as separate weights on Hugging Face. Technically, FLUX.1 uses a rectified flow transformer architecture (12 billion‑parameter variants are documented in model cards) and applies guidance / adversarial distillation techniques that enable high-quality outputs in very few sampling steps (the "schnell" variant is optimized for 1–4 step synthesis). Models are distributed both as open-weight dev variants and as commercial/pro tiers accessible via Black Forest Labs' API or licensing. Community adoption is measurable (the GitHub repo has tens of thousands of stars and multiple community ports, and the Hugging Face model pages show high monthly downloads), while community feedback ranges from praise for output quality and prompt‑adherence to questions about license gating and API support.

GitHub Statistics

• Stars: 25,020
• Forks: 1,836
• Contributors: 26
• License: Apache-2.0
• Primary Language: Python
• Last Updated: 2025-07-31T13:29:34Z

Key Features

• 12B rectified flow transformer architecture for high-fidelity text-to-image generation.
• FLUX.1 [schnell] produces high-quality images in 1–4 sampling steps via distillation.
• FLUX.1 Kontext enables multi-turn, locally consistent image editing from text instructions.
• Open-weight releases (Apache-2.0 for schnell) plus gated dev variants on Hugging Face.
• Integrations: Diffusers FluxPipeline, ComfyUI, community GGUF/quantized ports, and TensorRT support.

Example Usage

Example (python):

import torch
from diffusers import FluxPipeline

# Example: basic text-to-image with FLUX.1 [schnell]
model_id = "black-forest-labs/FLUX.1-schnell"

pipe = FluxPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16)
pipe = pipe.to("cuda")

prompt = "A photorealistic portrait of an astronaut cat wearing a vintage helmet"
image = pipe(prompt=prompt, num_inference_steps=2).images[0]

# Save output
image.save("flux_astronaut_cat.png")

Pricing

Black Forest Labs offers a self-serve commercial/dev licensing tier: $999/month (includes up to 100,000 images), then $0.01 per generated image above 100,000. Commercial licensing and higher-volume plans require contacting Black Forest Labs sales. (Source: BFL licensing and help pages.)

Benchmarks

Model parameter count: 12 billion parameters (FLUX.1 [dev]/[schnell] variants). (Source: https://huggingface.co/black-forest-labs/FLUX.1-schnell)

Generation steps (FLUX.1 [schnell]): High-quality images in 1–4 sampling steps via latent adversarial diffusion distillation. (Source: https://huggingface.co/black-forest-labs/FLUX.1-schnell)

Monthly downloads (FLUX.1 [dev]): 1,555,724 downloads (reported on Hugging Face model page, last-month metric). (Source: https://huggingface.co/black-forest-labs/FLUX.1-dev)

Monthly downloads (FLUX.1 [schnell]): 742,354 downloads (reported on Hugging Face model page, last-month metric). (Source: https://huggingface.co/black-forest-labs/FLUX.1-schnell)

Repository popularity / activity: GitHub: ~25k stars, ~1.8k forks, active issues and example tooling. (Source: https://github.com/black-forest-labs/flux)

Key Information

• Category: Image Models
• Type: AI Image Models Tool