Choose Auto, TokenLab Verified, or Official for each request, with prices shown up front. See what's new

Black Forest Labs: FLUX.2 [dev]

FLUX.2 dev is an open weight text to image and image editing model from Black Forest Labs. It targets developers who need precise control over prompts, references, and iteration. Use it for non commercial research, workflow prototyping, and multi conditioning image pipelines.
Compare models
flux-2-dev
AvailableBlack Forest LabsImageSync
Price
From $0.009
Modalities
Image

About FLUX.2 [dev]

FLUX.2 [dev] is Black Forest Labs' open-weight image model with 32 billion parameters, covering text-to-image generation, single-reference editing, and multi-reference editing in one checkpoint. It is the full-size, guidance-distilled member of the FLUX.2 family, aimed at developers who want precise prompt control and reference-driven pipelines rather than the sub-second speed of the smaller klein models.

Where it works well

  • Multi-reference editing: several input images plus a text prompt combine into one result, so characters, objects, and styles carry over without fine-tuning.
  • One model handles both generating from text alone and editing supplied images, which keeps a pipeline to a single integration.
  • At 32 billion parameters it is the largest open-weight FLUX.2 checkpoint, built for prompt fidelity over throughput.
  • Guidance distillation lets it reach usable quality without the extra guidance pass that undistilled diffusion models require.

When to choose another model

  • Black Forest Labs lists text rendering and factual accuracy as known weak points, so posters with long copy or diagrams need checking.
  • The smaller klein models return images far faster, which matters for interactive tools or high-volume batch jobs.
  • The weights carry a non-commercial license, so shipping the self-hosted checkpoint in a product needs a separate agreement with the maker.

Getting started

  1. Create API key

    Create a key in Console, then use it with every model on the platform.

  2. Send your first request

    Copy the example for your language and run it against the endpoint.

    Image editingText to imageTokenLab endpoint
    POST/v1/images/generations
    curl -X POST "https://api.tokenlab.sh/v1/images/generations" \
      -H "Authorization: Bearer sk-xxx" \
      -H "Content-Type: application/json" \
      -d '{
      "operation": "image-to-image",
      "model": "flux-2-dev",
      "prompt": "A minimalist product photo of matte black headphones on a soft blue background.",
      "image_url": "https://example.com/source.png"
    }'

Parameters

Image editing

ParameterValues and limitsDefaultNotes
cfg_scale≥ 1, ≤ 20Not providedUse multiples of 0.01.
guidance_scale≥ 1, ≤ 20Not providedUse multiples of 0.01.
height≥ 512, ≤ 20481024Use an integer. Use multiples of 16.
imageNot providedNot providedAlias of the combined images input. See input_limits.images for the total input image count.
image_urlNot providedNot providedAlias of the combined images input. See input_limits.images for the total input image count.
image_urlsNot providedNot providedAlias of the combined images input. See input_limits.images for the total input image count.
imagesInputs: ≥ 1, ≤ 4Not providedLimits apply when this input is supplied. Image aliases are combined into one input image list; the limit applies to the combined list.
n11Use an integer.
negative_promptNot providedNot providedMinimum length: 1 characters. Maximum length: 32000 characters.
operationRequiredNot providedNot providedNot provided
output_formatjpeg, png, webpNot providedDefault encoding is shared by these equivalent values: jpg, jpeg.
promptNot providedNot providedMinimum length: 1 characters. Maximum length: 32000 characters.
seed≥ 0, ≤ 9007199254740991Not providedUse an integer.
steps≥ 1, ≤ 5020Use an integer.
width≥ 512, ≤ 20481024Use an integer. Use multiples of 16.

Text to image

ParameterValues and limitsDefaultNotes
cfg_scale≥ 1, ≤ 20Not providedUse multiples of 0.01.
guidance_scale≥ 1, ≤ 20Not providedUse multiples of 0.01.
height≥ 512, ≤ 20481024Use an integer. Use multiples of 16.
imagesNot providedNot providedText-to-image does not accept source images. Do not send this field for that operation. Use one of this model's declared image operations for source-image requests.
n11Use an integer.
negative_promptNot providedNot providedMinimum length: 1 characters. Maximum length: 32000 characters.
operationRequiredNot providedNot providedNot provided
output_formatjpeg, png, webpNot providedDefault encoding is shared by these equivalent values: jpg, jpeg.
promptNot providedNot providedMinimum length: 1 characters. Maximum length: 32000 characters.
seed≥ 0, ≤ 9007199254740991Not providedUse an integer.
steps≥ 1, ≤ 5020Use an integer.
width≥ 512, ≤ 20481024Use an integer. Use multiples of 16.

Fields outside this contract are rejected.

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Image editing

per request
Official price
$0.009
Verified price
—
Discount
—

Text to image

per request
Official price
$0.009
Verified price
—
Discount
—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours
Requests
Success rate
P95 latency
Total tokens
Model performance
Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open FLUX.2 [dev] in Console with a prompt ready to edit or send.

Help me create an image with flux-2-dev using image-edit at /v1/images/edits. Show the result, cost, and any limits I should know.

Use cases

Best for
  • Text to Image
  • Reference-based product shots

    Feed a product photo and a style reference, then ask for the item in a new scene while keeping its shape and logo placement recognisable.

  • Character consistency across a series

    Give two or three reference images of one character and generate panels or storyboard frames where the face and outfit stay the same.

  • Workflow prototyping

    Try an image-editing step in a pipeline, then compare the output with the lighter klein models.

  • Prompt-heavy concept art

    Describe lighting, lens, composition, and materials in detail and iterate on the wording, since the model is tuned to follow long prompts.

Prompt examples

Place the ceramic mug from image 1 on the wooden desk from image 2, morning window light from the left, shallow depth of field.

Keep the woman from the reference photo exactly as she is, but change her jacket to a mustard yellow raincoat and set the scene on a rainy street at night.

A flat-lay of a hiking kit on moss: compass, enamel cup, folded map, 35mm film look, soft overcast light.

FAQ

What is FLUX.2 [dev]?

It is Black Forest Labs' 32-billion-parameter open-weight image model. One checkpoint does text-to-image, single-reference editing, and multi-reference editing. The company describes it as state of the art among open models for those three tasks.

Can FLUX.2 [dev] edit an existing image?

Yes. Supply an image with an instruction and it edits it, or supply several reference images and it merges their subjects and styles into a new picture. Black Forest Labs says this works without per-character or per-style fine-tuning.

How does FLUX.2 [dev] differ from the klein models?

Dev is the large model, tuned for prompt fidelity and reference handling. The klein models are 4B and 9B, step-distilled or base variants that trade some capacity for speed or trainability. Choose dev when quality of adherence matters more than turnaround.

Is the output of FLUX.2 [dev] marked as AI-generated?

Black Forest Labs states that the model applies pixel-layer watermarking and supports C2PA metadata for provenance. It also ships with content filters at inference, so some prompts may be refused or altered.

Does the license restrict FLUX.2 [dev]?

The open weights use the FLUX Non-Commercial License, which limits self-hosted use. Calling the model through an API is a separate matter from running the downloaded checkpoint, so read the license before deploying the weights yourself.

How much does FLUX.2 [dev] cost?

On TokenLab, FLUX.2 [dev] costs $0.009 per request. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

Which endpoint should FLUX.2 [dev] use?

Use https://api.tokenlab.sh/v1/images/edits for FLUX.2 [dev]. The request example below shows the matching code shape.

Which operations does FLUX.2 [dev] support?

FLUX.2 [dev] supports Image editing, Text to image. Select an operation above to see its endpoint and request example.

Compare FLUX.2 [dev]

Guides that use FLUX.2 [dev]

Sources

Reviewed Oct 2, 2026

More from FLUX

Related models