Image

FLUX.1 Kontext

flux-kontext-pro

FLUX.1 Kontext is a context-aware image model family for editing and text-to-image generation, with Pro and Max variants, predictable per-image pricing, and asynchronous task processing.

Context-Aware EditingText to ImageFixed Per-Image Pricing

Get API Key API Docs

Service Highlights

99.9% SLA
Official Discounts
Pay-as-you-go
Fast & Low Latency

Input

Model

Reference Image

Prompt

Aspect Ratio

Output Format

Seed (optional)

Prompt Upsampling

Safety Tolerance2

Output

Image link valid for 72 hours

Pricing Details

Transparent pricing with no hidden fees. Pay only for what you use.

* Actual costs are subject to final output.

FLUX.1 Kontext API Features

Editing-first capabilities with predictable controls and one asynchronous workflow

Context-Aware Image Editing

Use natural-language instructions to change a background, object, color, or visual treatment while the model considers the supplied image context.

Up to 4 Reference Images

Supply as many as four public image URLs for composition and identity cues. Multi-reference input is an experimental capability.

9 Aspect Ratios

Choose 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9, or 9:21 with the size field.

Format and Safety Controls

Choose PNG, JPEG, or WebP output and set safety_tolerance from 0 to 6.

Unified Asynchronous API

Submit with POST /v1/images/generations, then poll GET /v1/tasks/{task_id} until the task completes or fails.

Fixed Per-Image Billing

Each request returns one image at a fixed model price regardless of aspect ratio or reference count. Failed tasks are refunded automatically.

Pro and Max Options

Choose flux-kontext-pro for faster, lower-cost everyday work, or flux-kontext-max when output quality takes priority over speed and budget.

Prompt Enhancement and Repeatable Seeds

Enable prompt_upsampling to expand instructions, or reuse a fixed seed with the same inputs to help reproduce a result.

Where FLUX.1 Kontext Fits

Use a reference when visual context matters, or omit it for text-to-image generation

Products, Backgrounds & Objects

Create product variants, replace scenes, or modify selected objects through concise editing instructions.

Character & Style Continuity

Use an existing image as context to help carry character traits or visual style across new edits; results can vary by prompt and source image.

Experimental Multi-Reference Fusion

Combine cues from up to four images for compositions such as placing a subject into another scene or blending visual references.

Text-to-Image Generation

Send a prompt without image_urls to generate a new image with either Kontext variant.

Campaign and Social Assets

Edit a core visual into campaign or social-media variations, then select a supported aspect ratio for each placement.

Creative Concept Iteration

Start from an existing image, refine individual elements through text instructions, and explore controlled visual directions across iterations.

Start Using FLUX.1 Kontext

Create an account, authenticate, then submit and poll one unified image task

Create an account and add funds

→

Generate an API key

Create a key in the dashboard and send it as a Bearer token in the Authorization header.

→

Submit and poll the task

POST the request to /v1/images/generations, read data[0].task_id, and poll /v1/tasks/{task_id} until completed or failed.

Get API Key

FLUX.1 Kontext FAQ

Model choice, inputs, controls, billing, and task behavior

What is FLUX.1 Kontext?

FLUX.1 Kontext is Black Forest Labs' context-aware image model family. It is designed primarily for image editing from text instructions and can also generate images from text alone.

What is the difference between Pro and Max?

flux-kontext-pro is the faster, lower-cost option at the standard price of $0.04 per image and measured at about 6 seconds for one request. flux-kontext-max targets higher output quality at $0.08 per image and measured at about 9 seconds.

How is Kontext priced?

Each variant has a fixed standard price per generated image, independent of aspect ratio and reference-image count: $0.04 for Pro and $0.08 for Max. Your account's multiplier or discount may change the actual charge; check the task result for the recorded cost.

How many reference images can I send?

You can pass up to four URLs in image_urls. The output and all reference images must stay within 9MP in total. Multiple references are experimental, and every URL must be publicly accessible.

Can I request an exact image size?

Kontext outputs approximately 1MP, and resolution has no effect. width and height are rejected. Use size with one of the nine supported ratios or a WIDTHxHEIGHT pixel string; pixel strings are converted to the nearest supported ratio. Actual 16:9 and 9:16 outputs may differ from the requested ratio by about 4%.

Which formats and seed controls are available?

Use output_format to select png, jpeg, or webp. seed is optional and random by default; reusing a fixed seed with the same inputs can help reproduce a result.

Can Max rewrite my prompt?

It may. In measured calls, flux-kontext-max expanded the prompt even when prompt_upsampling was not explicitly enabled. Treat prompt rewriting as possible and review outputs when wording must remain strict.

How do asynchronous tasks, refunds, and public URLs work?

Submit a generation request and poll the returned task_id until completed or failed. Failed tasks are refunded automatically. Reference URLs must be publicly reachable; hosts that block data-center traffic can cause fetching to fail, so use a reliable public URL.

Related Models

Explore more models in the same category.

View All Models →