
01 · Product campaign visual
Make the cold-brew bottle the hero
Describe the product, light and negative space to create a key visual for brand campaigns, then check material detail and composition.
FLUX 3 Image by Black Forest Labs is the image model of the FLUX 3 API. It generates one image per request from text or up to 10 references, at five resolution tiers up to 4K. Resolution sets the price; reference images add no separate charge.

Transparent pricing with no hidden fees. Pay only for what you use.
Featured examples
Pick an example to load the model, prompt and full settings, then adjust it your way.

01 · Product campaign visual
Describe the product, light and negative space to create a key visual for brand campaigns, then check material detail and composition.
Reference02 · Multi-image compositing
Load the person and the location in order, keep her look and outfit, and blend her into the desert rock scene, then check the light direction and shadows. See the edited result in the output area.
Reference03 · Style transfer
Load the reference image and keep the street layout, doors, windows and distant harbor while switching to fine ink linework with transparent watercolor. See the edited result in the output area.
Use the FLUX 3 API to create images from text, edit a reference or combine up to 10 images. Choose a resolution and aspect ratio, control search grounding, and describe layouts in your prompt. Klein and Video use the same APIMart key.
FLUX 3 Image Klein is the smaller, faster open-weights model in the same family. It adds seed, 1–16 steps and PNG or JPEG output at a lower price per image. View FLUX 3 Image Klein
Generation, editing and output controls supported by the API
Describe the subject, composition, lighting and style to create a new image.
Edit one image or combine up to 10 references with instructions in the prompt.
Choose 768sq, 1k, 1.5k, 2k or 4k. The default is 1k; higher resolutions affect the price.
Choose from 15 ratios or auto. When editing, auto follows the first reference image.
Describe bounding-box layout and local edits in the prompt using the documented syntax. This is not a visual box-drawing editor.
Allow the upstream model to retrieve web or image context before generation. Grounding is on by default and can be disabled.
Set safety_tolerance from 0 to 4, with 0 being strictest. The default is 2.
Each request generates one image. Price depends on resolution, not aspect ratio or the number of references.
Suggested uses for creative, commerce, and content teams
Explore product-scene drafts and local adjustments, with checks for product details and faithful presentation before use.
Develop campaign directions and plan element placement, then review the selected image against the creative brief.
Create still-image drafts to communicate a scene or presentation concept; review each image before assembling a sequence.
Try targeted changes to a design and compare them with earlier drafts, keeping human approval in the workflow.
Develop illustration concepts for articles and guides, with editorial review to keep imagery aligned with the story.
Explore composition ideas for packaging or posters, reserving space for copy and checking final brand and print requirements separately.
Three steps from sign-up to your first FLUX 3 Image API request
Sign up for free. One account gives you the FLUX 3 Image API and every other model on APIMart.
Top up your balance. FLUX 3 Image is billed per image by resolution tier, and the same balance works across all models.
Create an API key in the dashboard and send it as a Bearer token to POST /v1/images/generations with model flux-3-image.
Call flux-3-image with one POST request. Generation is asynchronous: submit a task, then poll the task endpoint for the image URL.
Save your APIMart key in the APIMART_API_KEY environment variable and send it as Authorization: Bearer <key> on every request.
POST /v1/images/generations with model flux-3-image, your prompt, resolution and aspect_ratio. Add image_urls to edit or combine references. The example.com URLs are placeholders: replace them with your own publicly accessible images or Base64 data. The task ID is in data[0].task_id.
Poll GET /v1/tasks/{task_id} until status is completed, then read the image URL from data.result.images[0].url[0]. APIMart keeps results for 72 hours, so download the images you want to keep.
See the FLUX 3 Image API reference for every parameter, layout syntax and response field.
curl -X POST https://api.apimart.ai/v1/images/generations \
-H "Authorization: Bearer $APIMART_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "flux-3-image",
"prompt": "A ceramic coffee cup on a linen tablecloth by a window, soft morning light, product photo, the word MORNING printed on the cup",
"resolution": "1k",
"aspect_ratio": "16:9"
}'import os
import time
import requests
API_BASE = "https://api.apimart.ai"
headers = {"Authorization": f"Bearer {os.environ['APIMART_API_KEY']}"}
# 1. Submit the generation task
payload = {
"model": "flux-3-image",
"prompt": "A ceramic coffee cup on a linen tablecloth by a window, soft morning light, product photo, the word MORNING printed on the cup",
"resolution": "1k",
"aspect_ratio": "16:9"
}
resp = requests.post(f"{API_BASE}/v1/images/generations", headers=headers, json=payload, timeout=60)
resp.raise_for_status()
task_id = resp.json()["data"][0]["task_id"]
# 2. Poll the task until it finishes
for _ in range(200):
task_resp = requests.get(f"{API_BASE}/v1/tasks/{task_id}", headers=headers, timeout=30)
task_resp.raise_for_status()
task = task_resp.json()["data"]
if task["status"] == "completed":
print(task["result"]["images"][0]["url"][0])
break
if task["status"] == "failed":
raise RuntimeError(task.get("error", {}).get("message", "generation failed"))
time.sleep(3)
else:
raise TimeoutError(f"task {task_id} did not finish in time; continue GET polling with this task ID")curl -X POST https://api.apimart.ai/v1/images/generations \
-H "Authorization: Bearer $APIMART_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "flux-3-image",
"prompt": "Replace the background with a sunny beach and keep the product, its label and its lighting unchanged",
"resolution": "1k",
"aspect_ratio": "auto",
"image_urls": [
"https://example.com/reference.jpg"
]
}'TASK_ID="task_xxx" # replace with the task_id returned by the POST request
curl "https://api.apimart.ai/v1/tasks/$TASK_ID" \
-H "Authorization: Bearer $APIMART_API_KEY"| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
model | string | Required | Default: flux-3-image | Use flux-3-image. |
prompt | string | Required | Default: — | Scene description for text-to-image, or edit instructions. Negative prompts are not supported; describe what you want to see. |
image_urls | string[] | Optional | Default: — | Up to 10 reference images as public HTTP(S) URLs or Base64. Omit for text-to-image; the first image is the one being edited. |
resolution | string | Optional | Default: 1k | 768sq, 1k, 1.5k, 2k or 4k (about 16 MP). Sets the price per image. |
aspect_ratio | string | Optional | Default: auto | One of 15 ratios from 21:9 to 9:21, or auto. Pixel sizes such as 1024x1024 return HTTP 400. |
grounding | boolean | Optional | Default: true | Allow web or image search before generation. Send the boolean false to disable it. |
safety_tolerance | integer | Optional | Default: 2 | 0 (strictest) to 4. |
One APIMart API key calls every FLUX model below. Image models are billed per image and FLUX 3 Video per second of video.
| Model | Best for | Output |
|---|---|---|
| FLUX 3 Image APIThis page | The FLUX 3 API for images: generation and editing with up to 10 references, search grounding and output up to 4K | 768sq to 4K (about 16 MP) |
| FLUX 3 Image Klein API | Lower-cost FLUX 3 images with seed, 1–16 steps and PNG or JPEG output | 768sq to 4K (about 16 MP) |
| FLUX 3 Video API | Text to video, 1–10 keyframes, video continuation and drafts, with synced audio | HD or FHD, 5–20 s |
| FLUX.2 API | Choosing between FLUX.2 Pro, Max and Flex on one page | 1–4 MP |
| FLUX.2 Pro API | Balanced quality and cost for production images, up to 8 references | 1–4 MP |
| FLUX.2 Max API | The highest-quality FLUX.2 output, up to 8 references | 1–4 MP |
| FLUX.2 Flex API | Adjustable steps and guidance for typography and layout-heavy designs | 1–4 MP |
| FLUX.1 Kontext API | Choosing between FLUX.1 Kontext Pro and Max on one page | About 1 MP |
| FLUX.1 Kontext Pro API | Context-aware edits at a fixed price per image, up to 4 references | About 1 MP |
| FLUX.1 Kontext Max API | Higher-quality Kontext edits with stronger prompt adherence | About 1 MP |
Inputs, output settings, billing and task handling
FLUX 3 is Black Forest Labs' multimodal model family. On APIMart the FLUX 3 API covers three models with one key: FLUX 3 Image (flux-3-image, this page) for image generation and editing, FLUX 3 Image Klein (flux-3-image-klein) for lower-cost images, and FLUX 3 Video (flux-3-video) for video with audio. Each FLUX 3 Image request returns one final image.
Black Forest Labs announced FLUX 3 on July 23, 2026 and released FLUX 3 Image through its API on October 1, 2026, according to its release notes. On APIMart you call it with model flux-3-image.
image_urls accepts up to 10 publicly accessible HTTP(S) URLs or base64 images. Adding references makes the request an editing or multi-reference task. References do not add a separate charge.
Resolutions are 768sq, 1k, 1.5k, 2k and 4k. Choose auto or one of 15 ratios from 21:9 to 9:21. Defaults are 1k and auto. Auto follows the first reference when editing; text-only generation follows the prompt or defaults to 1:1.
grounding defaults to true and allows upstream web or image retrieval before generation; set it to false to disable retrieval. safety_tolerance is an integer from 0 to 4, where 0 is strictest and the default is 2.
Layout and local-edit instructions belong in prompt, using the syntax described in the official documentation. The tool does not provide a visual bounding-box editor or a separate bbox parameter.
The price per image depends only on resolution. Original prices on APIMart (October 2026): 768sq $0.041, 1k $0.048, 1.5k $0.07, 2k $0.10 and 4k $0.607. Aspect ratio and reference count do not change the price. Gold members pay 20% less and Platinum and Diamond members save up to 28%; the pricing table on this page shows current rates, and the completed task's cost is the final charge. Failed or moderation-blocked tasks are fully refunded.
POST /v1/images/generations returns a task_id. Query GET /v1/tasks/{task_id} until the task completes or fails, then read the image URL from data.result.images[0].url[0]. A 4k task may take several minutes. APIMart keeps results for 72 hours, so download the images you want to keep.
Not without checking. This API does not accept width, height, pixel-size strings, seed, steps, guidance, output_format, negative_prompt or prompt_upsampling, and it does not generate multiple images per request. Choose resolution and aspect_ratio instead.
Sign up for APIMart, add credits, then create a key on the API Keys page of the dashboard. Send it as a Bearer token with each request. The same key calls FLUX 3 Image, FLUX 3 Image Klein, FLUX 3 Video and every other model on APIMart, with pay-as-you-go billing.
Explore more models in the same category.

Nano Banana 2.1
Nano Banana 2.1 is Google's Gemini image model for text-to-image generation and reference-image editing. It outputs 1K, 2K or 4K images in 14 aspect ratios and is available through official and Ext routes.

Seedream 5.0 Flash
Seedream 5.0 Flash is ByteDance's fast, lower-cost text-to-image model. It supports single-image and multi-reference editing with up to 10 references and generates 1K, 1.5K or 2K images, or exact pixel sizes, with strong text rendering.

GPT Image 2.5 Flare
GPT Image 2.5 Flare is OpenAI's speed-focused GPT Image 2.5 model for image generation and editing. It supports up to 16 reference images, inpainting with transparent PNG masks and 15 aspect ratios.

GPT Image 2.5 Sunburst
GPT Image 2.5 Sunburst is an OpenAI image model for generation and editing, focused on precise edits for detail-sensitive creative work. Supports up to 16 reference images and PNG, JPEG or WebP output.
Guides, comparisons and pricing analysis from the APIMart blog.