1. Tutorial Introduction

FLUX.2 is an AI image model launched by Black Forest Labs in November 2025. It is designed specifically for real-world creative workflows. The model supports multi-image references of up to 10 images, generating high-quality images up to 4MP resolution with exceptional detail and text rendering capabilities. Combining a visual language model with a stream transformer architecture, the model significantly improves real-world knowledge understanding and image generation quality, driving open innovation and widespread application of visual intelligence technology.

This tutorial uses a single RTX PRO 6000 GPU for computing power, and deploys a 4-bit quantized model: diffuses/FLUX.2-dev-bnb-4bit. Two examples are provided for testing: multi-reference editing and text-to-image generation.

2. Effect display

multi-reference editing

text-to-image generation

3. Operation steps

1. Start the container

2. Usage steps

If "Bad Gateway" is displayed, it means the model is initializing. Since the model is large, please wait about 2-3 minutes and refresh the page.

1. Multi-reference editing

Specific parameters:

• Seed: An initial value input into the random number generator of the generative model, used to control the randomness in the generation process.
• Width: The width of the generated image.
• Height: The height of the generated image.
• Number of inference steps: refers to the number of iterations or processing steps that the generative model undergoes in generating the final result.
• Guidance scale: controls the degree of influence of conditional inputs on the final generated result in generative models (such as diffusion models).

2. Text-to-image generation

Specific parameters:

• Seed: An initial value input into the random number generator of the generative model, used to control the randomness in the generation process.
• Width: The width of the generated image.
• Height: The height of the generated image.
• Number of inference steps: refers to the number of iterations or processing steps that the generative model undergoes in generating the final result.
• Guidance scale: controls the degree of influence of conditional inputs on the final generated result in generative models (such as diffusion models).

Citation Information

The citation information for this project is as follows:

@misc{flux-2-2025,
    author={Black Forest Labs},
    title={{FLUX.2: Frontier Visual Intelligence}},
    year={2025},
    howpublished={\url{https://bfl.ai/blog/flux-2}},
}