BLACK FOREST LABS · AI IMAGE GENERATOR
FLUX 3 Image is the image part of Black Forest Labs' newest model family. Describe a scene, edit a photo with one sentence, or combine up to 10 reference images — then pick 1K, 1.5K, 2K or 4K output.
Review results, manage assets and access generation history in AI Studio.
Every example below was actually generated with FLUX 3 on ImgAI, with the original prompt and settings. Click "Use this example" to load it into the creator.
FLUX 3 · 1.5K · 2:3
Both lines — "SUMMER LIGHT" and "JULY 18–20 · LAKESIDE PARK" — came out exactly as written, above a flat orange sun setting into stylized waves on paper texture.
A minimalist summer music festival poster. Large bold serif title "SUMMER LIGHT" at the top, smaller line "JULY 18–20 · LAKESIDE PARK" below it, a stylized orange sun setting over calm water in the center, off-white paper texture, generous margins.
FLUX 3 · 1K · 1:1
A glass perfume bottle with a brushed gold cap on a travertine block, with a dried palm leaf and its shadow behind — soft light and believable glass.
Studio product photo of a faceted glass perfume bottle with a brushed gold cap on a pale travertine block, soft window light from the left, a single dried palm leaf casting a long shadow, clean beige background.
FLUX 3 · 1K · 16:9
Each element was given a box in the prompt: the celadon teapot sits on the left, two cups on the right and the jasmine sprig between them, as placed.
Editorial magazine layout photo: a ceramic teapot <teapot_1> on the left, two matching cups <cups_1> on the right, and a small sprig of jasmine <jasmine_1> between them on a linen tablecloth <table_1>. [{"id":"table_1","bbox":[550,0,1000,1000],"desc":"cream linen tablecloth with soft folds"},{"id":"teapot_1","bbox":[250,80,800,430],"desc":"celadon green ceramic teapot"},{"id":"cups_1","bbox":[480,600,800,940],"desc":"two small celadon cups"},{"id":"jasmine_1","bbox":[620,450,780,560],"desc":"a sprig of white jasmine flowers"}]
FLUX 3 · 1K · 4:3
One sentence added a sleeping ginger cat on the window seat; the shelves, lamp, side table and light match the original photo.
Add a ginger cat curled up asleep on the window seat cushion. Keep the shelves, books, window, lighting and camera angle exactly the same.
FLUX 3 · 1K · 4:3
The wristwatch from the first reference now stands on a shelf of the reading nook from the second, small and in scale; the rest of the room is unchanged.
Place the wristwatch from Image 1 standing on the middle oak shelf of the reading nook in Image 2, scaled naturally, lit by the same soft morning light. Keep the rest of Image 2 unchanged.
FLUX 3 · 2K · 21:9
A 21:9 frame at 2K: a vintage car with headlights on rounds a cliffside bend above breaking waves under a pastel dawn sky.
Ultra-wide cinematic shot of a misty coastal highway at dawn, a vintage car with headlights on rounding a cliffside bend, waves breaking far below, pastel sky.
FLUX 3 is Black Forest Labs' multimodal model family for images, video, audio and actions; FLUX 3 Image is the part that generates and edits images. BFL highlights control: you can place elements with boxes on a 0–1000 grid, revise one area and leave the rest of the picture as it was, and combine up to 10 references.
On ImgAI you choose from 15 aspect ratios and four sizes (1K, 1.5K, 2K, 4K). Reference images are free to add — the price depends only on the output size: 55–695 credits per image. Optional web search can look up real-world context before drawing.
Based on the model's published capabilities and what we actually generated, to help you judge whether it fits your task.
Titles, dates and images arranged where you ask, with clean type.
Glass, metal and soft studio light rendered convincingly.
Add or change one thing and keep the rest of the shot.
Put a product from one image into the scene of another.
Specs come from the configuration currently available on this site; the creator always shows the live price. Failed generations are refunded automatically.
| Input | Text + up to 10 reference images |
| Resolution | 1K · 1.5K · 2K · 4K |
| Prompt length limit | 10,000 characters |
Supported aspect ratios
| Tier | 1K | 1.5K | 2K | 4K |
|---|---|---|---|---|
| FLUX 3 | 55 | 80 | 115 | 695 |
These come from real generations — read them alongside the original prompts in the examples.
For edits, name what to change and what to keep: "keep the shelves, light and angle the same".
With several references, write "Image 1", "Image 2" in the order you uploaded them.
Tag elements like <teapot_1> and append a JSON list of boxes ([top, left, bottom, right], 0–1000). See the layout example below.
1K for drafts and social posts, 2K for web pages, 4K when you will crop or print — 4K costs much more.
It is the image model in Black Forest Labs' FLUX 3 family. It does text-to-image, single-image edits and multi-reference composition in one model.
55–695 credits per image, depending only on output size. Adding reference images or web search does not change the price. Failed generations are refunded automatically.
Up to 10. One image works as a single edit; several images can be combined, referred to as Image 1, Image 2 and so on.
A way to tell the model where each element goes. Coordinates run from 0 to 1000 on both axes in [top, left, bottom, right] order and are written as JSON at the end of the prompt. BFL describes boxes as placement guides, not hard masks.
Yes. The provider's own documentation uses Chinese example prompts, and our tests with Chinese prompts followed the description closely.
On ImgAI FLUX.2 is text-to-image only and Kontext takes up to 4 references at about 1MP. FLUX 3 does both in one model, with up to 10 references and output up to 4K.
Sign up and generate with credits. Results are saved to your library, ready to download or keep editing.