Flux
What is Flux?
Flux is an AI image generation and editing platform for product teams, creators, developers, and enterprises that need controllable visuals at production scale. It combines text-to-image generation, local editing, inpainting/outpainting, and multi-reference control, with iterative edits that preserve character consistency. Flux also offers self-hosting and integrations with Hugging Face and GitHub. Plans are custom across FLUX.2 [flex], [pro], [max], [klein] 4B, [klein] 9B, and FLUX.1 [dev].
Last verifiedHow we evaluate
At a glance
- Flux is best for product teams and creators who need controllable image generation and editing at production scale.
- FLUX.2 [flex] Custom; FLUX.2 [pro] Custom; FLUX.2 [max] Custom; FLUX.2 [klein] 4B Custom; FLUX.2 [klein] 9B Custom; FLUX.1 [dev] Custom
What does Flux do?
Flux handles production-grade image generation and editing through a browser playground, API, and open-weights models. Its workflow combines text-to-image generation, local editing, generative modifications, and inpainting/outpainting, so teams can create new visuals or transform existing ones without switching tools. The site shows multi-reference control and iterative editing that preserves character consistency across multiple turns, plus a "raw" option for extra realism on photo-realistic outputs. At scale, Flux is built for production workloads and self-hosted deployment, with sub-second inference time on capable hardware and 4MP photorealistic output. Customers shown on the site include Adobe, Microsoft, Mistral, Freepik, Picsart, Gamma, OpenArt, and Burda Verlag.
Why use Flux?
- Multi-reference control helps teams keep characters and visual details consistent across repeated edits.
- Open-weights access lets organizations run FLUX on their own infrastructure instead of depending only on hosted access.
- The platform combines generation, editing, and inpainting/outpainting in one workflow, reducing tool switching.
- Sub-second inference on capable hardware supports interactive iteration instead of batch-only image generation.
- The browser playground lets teams test ideas instantly before wiring the models into production.
Who is Flux for?
- Product teams that need consistent image generation across iterative editing workflows.
- Design and content teams that want photorealistic visuals with precise regional edits.
- Developers building visual AI features through an API or self-hosted deployment.
- Enterprises that need deployment flexibility, data sovereignty, and compliance controls.
- Creative platforms that need multi-reference generation for production workloads.
What are Flux's key features?
Production-grade AI image generation and editing
Generate and edit images with FLUX.2 models, including text-to-image, local editing, and inpainting/outpainting with binary masks for precise scene changes.
multi-reference control
Guide outputs with multiple reference images to keep character consistency across iterative edits, while supporting prompt adherence and typography generation.
Deploy anywhere
Run the models through self-hosting or integrate with Hugging Face and GitHub, giving teams control over deployment and workflow fit.
Customizable
Tune image generation for editing, generative modifications, and large-scale workloads, with support for 4MP photorealistic output and output resizing rules.
Try instantly
Test the models quickly with sub-second inference time on capable hardware, making it easier to validate quality before building a workflow.
No code required
Use the hosted experience to create and edit images without building a custom pipeline, while still working with text and image inputs.
Explore and experiment
Experiment with iterative workflows across multiple editing turns, comparing model behavior on prompt adherence, typography, and realism before production use.
What does Flux integrate with?
- Hugging Face
- GitHub
What are Flux's use cases?
Iterative editing for product teams
Product teams use Flux to keep characters, layouts, and style consistent across repeated image revisions, using Production-grade AI image generation and editing to move from concept to polished assets without restarting each round. They can refine scenes quickly while preserving the same visual direction.
Regional edits for design teams
Design and content teams use Flux to make precise local changes to photorealistic visuals, using multi-reference control to guide composition and Full model access to choose the right model for the job. That helps them produce campaign-ready images with fewer manual touch-ups.
API visuals for developers
Developers building visual AI features use Flux through an API or self-hosted setup, using Deploy anywhere to fit their infrastructure and Customizable to adapt outputs to product requirements. They can ship image generation features while keeping control over deployment and behavior.
Production workflows for creative platforms
Creative platforms use Flux to generate and edit images at scale, using multi-reference control to keep outputs aligned across many assets and Production-grade AI image generation and editing to support production workloads. That makes it easier to deliver consistent visuals for users and clients.
How does Flux work?
- Connect your first workflow through Try instantly or a supported integration like Hugging Face or GitHub, then open the model playground to test prompts and image inputs right away.
- Choose a model with Full model access, then use Explore and experiment to compare generation, editing, and regional change behavior before you commit to a production path.
- Add reference images and enable multi-reference control to steer composition, character consistency, and style across iterative edits, especially when you need repeatable visual results.
- Tune outputs with Customizable settings, then move into Production-grade AI image generation and editing for photorealistic assets, inpainting, outpainting, or full scene transformations.
- Deploy anywhere when you need tighter infrastructure control, or keep iterating in the hosted flow until your team is ready to scale visual generation into production.
How much does Flux cost?
FLUX.2 [flex]
Custom- Our new model brings maximum performance across all aspects, greatly improved prompt adherence and typography generation meet consistency for editing without compromise on speed.
- A unified model delivering local editing, generative modifications and text-to-image generation in FLUX.1 quality. Processes text and image inputs for precise regional edits or full scene transformations at breakthrough speeds, pioneering iterative workflows that maintain character consistency across multiple editing turns.
- Our best model for photo-realistic images at 2k resolution, with a "raw" option for extra realism.
- Our original pro model.
- Our distilled model of FLUX.1 [pro].
- Generated image: We charge $0.10 for each output megapixel. Output megapixel are defined by the height and width parameter in the API request. Input megapixels are not counted.
- Important Notes:
- For pricing, resolution is always rounded up to the next megapixel for the generated image.
- Output megapixels are determined by the output width and height
- Images exceeding 4 megapixels are resized to 4 megapixels
- Flux.1 based inpainting and outpainting. Modify an image given a text prompt and a binary mask.
FLUX.2 [pro]
Custom- Our new model brings maximum performance across all aspects, greatly improved prompt adherence and typography generation meet consistency for editing without compromise on speed.
- A unified model delivering local editing, generative modifications and text-to-image generation in FLUX.1 quality. Processes text and image inputs for precise regional edits or full scene transformations at breakthrough speeds, pioneering iterative workflows that maintain character consistency across multiple editing turns.
- Our best model for photo-realistic images at 2k resolution, with a "raw" option for extra realism.
- Our original pro model.
- Our distilled model of FLUX.1 [pro].
FLUX.2 [max]
Custom- Our new model brings maximum performance across all aspects, greatly improved prompt adherence and typography generation meet consistency for editing without compromise on speed.
- A unified model delivering local editing, generative modifications and text-to-image generation in FLUX.1 quality. Processes text and image inputs for precise regional edits or full scene transformations at breakthrough speeds, pioneering iterative workflows that maintain character consistency across multiple editing turns.
- Our best model for photo-realistic images at 2k resolution, with a "raw" option for extra realism.
- Our original pro model.
- Our distilled model of FLUX.1 [pro].
- Generated image: We charge $0.10 for each output megapixel. Output megapixel are defined by the height and width parameter in the API request. Input megapixels are not counted.
- Important Notes:
- For pricing, resolution is always rounded up to the next megapixel for the generated image.
- Output megapixels are determined by the output width and height
- Images exceeding 4 megapixels are resized to 4 megapixels
FLUX.2 [klein] 4B
Custom- Our new model brings maximum performance across all aspects, greatly improved prompt adherence and typography generation meet consistency for editing without compromise on speed.
- A unified model delivering local editing, generative modifications and text-to-image generation in FLUX.1 quality. Processes text and image inputs for precise regional edits or full scene transformations at breakthrough speeds, pioneering iterative workflows that maintain character consistency across multiple editing turns.
- Our best model for photo-realistic images at 2k resolution, with a "raw" option for extra realism.
- Our original pro model.
- Our distilled model of FLUX.1 [pro].
FLUX.2 [klein] 9B
Custom- Our new model brings maximum performance across all aspects, greatly improved prompt adherence and typography generation meet consistency for editing without compromise on speed.
- A unified model delivering local editing, generative modifications and text-to-image generation in FLUX.1 quality. Processes text and image inputs for precise regional edits or full scene transformations at breakthrough speeds, pioneering iterative workflows that maintain character consistency across multiple editing turns.
- Our best model for photo-realistic images at 2k resolution, with a "raw" option for extra realism.
- Our original pro model.
- Our distilled model of FLUX.1 [pro].
- Generated image: We charge $0.10 for each output megapixel. Output megapixel are defined by the height and width parameter in the API request. Input megapixels are not counted.
- Important Notes:
- For pricing, resolution is always rounded up to the next megapixel for the generated image.
- Output megapixels are determined by the output width and height
- Images exceeding 4 megapixels are resized to 4 megapixels
- Flux.1 based inpainting and outpainting. Modify an image given a text prompt and a binary mask.
FLUX.1 [dev]
Custom- Our new model brings maximum performance across all aspects, greatly improved prompt adherence and typography generation meet consistency for editing without compromise on speed.
- A unified model delivering local editing, generative modifications and text-to-image generation in FLUX.1 quality. Processes text and image inputs for precise regional edits or full scene transformations at breakthrough speeds, pioneering iterative workflows that maintain character consistency across multiple editing turns.
- Our best model for photo-realistic images at 2k resolution, with a "raw" option for extra realism.
- Our original pro model.
- Our distilled model of FLUX.1 [pro].
- Flux.1 based inpainting and outpainting. Modify an image given a text prompt and a binary mask.
Frequently asked questions
What is Flux?
Flux is an AI image generation and editing platform for product teams, creators, developers, and enterprises that need controllable visuals at production scale. It combines text-to-image generation, local editing, inpainting/outpainting, and multi-reference control, with iterative edits that preserve character consistency. Flux also offers self-hosting and integrations with Hugging Face and GitHub. Plans are custom across FLUX.2 [flex], [pro], [max], [klein] 4B, [klein] 9B, and FLUX.1 [dev].
How much does Flux cost? Is it free?
Flux has 6 paid plans: FLUX.2 [flex] at Custom, FLUX.2 [pro] at Custom, FLUX.2 [max] at Custom.
What is Flux used for? Who is it for?
Flux is used for Production-grade AI image generation and editing, multi-reference control, and Full model access. It's built for Product teams that need consistent image generation across iterative editing workflows, Design and content teams that want photorealistic visuals with precise regional edits, and Developers building visual AI features through an API or self-hosted deployment.
Does Flux have an API and what does it integrate with?
Flux doesn't publish a public API. It integrates with Hugging Face, GitHub.
Editor's read
Check whether your workflow needs the 4MP ceiling and output-megapixel billing on FLUX.2 [flex], [max], and [klein] 9B. If you plan to generate larger images, confirm how the resize-to-4MP rule affects cost and output quality before committing.
Explore other Design Agents
