Black Forest Labs makes the FLUX family of models and sells access as pay-as-you-go API credits, with no subscription plan and no free tier. What decides whether it suits you is whether you want a metered API and open-weights self-hosting rather than a monthly plan, and whether you need the commercial rights and ownership that come with every generation.

What is Black Forest Labs?
Black Forest Labs is a United States company, with its legal entity based in San Francisco. It runs as a browser playground and an API. Black Forest Labs builds its own FLUX models and does not offer other companies' models.
What Black Forest Labs generates
The newest model is FLUX 3, one multimodal model spanning image, video, audio and action prediction. FLUX.2 models output up to 4 megapixels, close to 2048 by 2048 pixels. FLUX.2 [flex] is specialized for typography and text rendering inside an image, and character and style consistency are covered by a dedicated guide for maintaining identity across edits.
Generation is steered with a fixed seed, prompt upsampling, and pose or layout guidance drawn from references like poses, depth maps or edge outlines. Most FLUX models do not support negative prompts. Editing tools include multi-reference image editing across up to 10 images and outpainting for aspect-ratio changes. Style and LoRA training is documented for FLUX.2 [klein], and every open-weights license includes fine-tuning rights. There is no background-remover tool and no vector output. An API gives programmatic access to all of it, with batch requests multiplying the per-image cost by the number of images.
Working with other people
Black Forest Labs runs on the web through its playground and API only, with no desktop or mobile app. Enterprise adds single sign-on through Okta, Azure AD, Google Workspace or another SAML or OIDC provider, and a managed API with zero data retention.
Black Forest Labs pricing: what each plan gives

Black Forest Labs sells pay-as-you-go credits, not a subscription: 1 credit costs $0.01, and there is no free tier and no free trial. FLUX.1 Kontext [pro] and FLUX1.1 [pro] cost 4 credits, $0.04, per image; batch requests multiply this by the number of images. Open-weights self-hosting is sold separately as annual or contact-sales licenses, without a listed price on the pricing page.
Rights and privacy
Using FLUX through the API includes full commercial rights with no separate license, and you own all right, title and interest in the output. Black Forest Labs' terms have you indemnify the company, not the other way round.
Your data is used for training by default; you can opt out by emailing privacy@blackforestlabs.ai with the subject Training Opt Out. Zero data retention is available on the Enterprise managed API. Black Forest Labs holds a SOC 2 Type II audit, processes data in a GDPR-compliant way as standard, and offers single sign-on. Support runs through email, live chat and a help center.
Who Black Forest Labs suits
Black Forest Labs suits developers who want metered, pay-as-you-go access to FLUX with full commercial rights and output ownership built in, or teams that want to self-host open-weights models, including the Apache 2.0 licensed FLUX.2 [klein] 4B, on their own infrastructure.
It suits less well anyone who wants a free tier to try it first, since none exists, or anyone who wants a negative prompt, since most FLUX models do not support one.
How Black Forest Labs compares
Black Forest Labs scores 7.2 out of 10 among the AI image generators we rate. These three do the same job and are the closest to it, compared on what each vendor publishes.
Midjourney
Black Forest Labs runs on pay-as-you-go credits with no monthly plan, while Midjourney's cheapest tier is a flat $10 a month; neither offers a free plan. Black Forest Labs publishes open weights under a named license, lets you self-host, and takes pose, sketch or depth map input, none of which Midjourney has. Midjourney answers with negative prompts to say what to leave out, a feature missing from Black Forest Labs' API.
Pick Midjourney if you want negative prompts and a flat monthly price instead of credits.
Pick Black Forest Labs if you want open weights, self-hosting or pose and depth control.
Stability AI
Black Forest Labs keeps a character the same across images, a feature Stability AI's API does not have. Stability AI answers with one-action background removal and negative prompts to say what to leave out, both missing from Black Forest Labs. Black Forest Labs generates a batch in one run and rewrites your prompt for you, while Stability AI's API returns one image per request with no prompt rewriting.
Pick Stability AI if you want background removal and negative prompts over character consistency.
Pick Black Forest Labs if you want character consistency, batch runs or automatic prompt rewriting.
Adobe Firefly
Firefly has a free plan; Black Forest Labs sells pay-as-you-go credits with no plan and no free trial. Firefly's terms say it does not train on what you make by default, while Black Forest Labs trains unless you opt out. Black Forest Labs publishes open weights and lets you self-host, and Firefly answers with one-action background removal and true vector exports instead.
Pick Adobe Firefly if you want a free plan and training off by default.
Pick Black Forest Labs if you want open weights, self-hosting or pay-as-you-go pricing.
Frequently Asked Questions
How much does Black Forest Labs cost?
There is no subscription and no free tier. Credits cost $0.01 each, and an image on FLUX.1 Kontext [pro] or FLUX1.1 [pro] costs 4 credits, $0.04.
Can I self-host a FLUX model?
Yes. Open-weights models, including FLUX.1 [dev], FLUX.2 [dev] and FLUX.2 [klein] at 4B and 9B, can be deployed on your own infrastructure, with the 4B [klein] model under an Apache 2.0 license.
Do I own what FLUX generates?
Yes. You own all right, title and interest in the output, and API usage includes full commercial rights with no separate license.
Does Black Forest Labs train on my data?
By default, yes, though you can opt out by emailing privacy@blackforestlabs.ai with the subject Training Opt Out.











