FLUX API Pricing: Full Breakdown of Costs (Jun 2026)

This guide covers what the FLUX (Black Forest Labs) API costs, what drives the price up, how to bring it down, and how to use FLUX for free, with a companion walkthrough in our free FLUX API tutorial.

How much does the FLUX API cost?

FLUX.2 [max], the top model in the lineup, starts at $0.07 per image. The cheapest option is FLUX.2 [klein] 4B at $0.014 per image. FLUX is an image generation API, so you pay per image rather than per token, and on the FLUX.2 family the price scales with the resolution you generate.

One caveat before you budget from this table. FLUX.2 prices are listed as "from" because they rise with output resolution, so a single 4MP image costs more than the headline figure.

How FLUX API pricing works

FLUX bills per generated image. There are two pricing models depending on the model family.

Megapixel-based pricing (FLUX.2)

The FLUX.2 family charges by megapixel of output. One megapixel is one million pixels, so a 1024×1024 image is about one megapixel. The first megapixel is a flat rate, and each additional megapixel adds a smaller increment. With FLUX.2 [klein] 4B, the first megapixel costs $0.014 and each additional megapixel adds $0.001, so a 2MP image costs $0.015. This is why every FLUX.2 price is quoted as "from": the figure is the one-megapixel floor, and your real cost depends on the width and height you request.

Flat credits per image (FLUX.1 and Kontext)

The previous-generation FLUX.1 models, including the Kontext editing models, use flat credit pricing instead of megapixels. One credit equals $0.01. FLUX1.1 [pro] is 4 credits ($0.04) per image, Kontext [pro] is 4 credits ($0.04), and Kontext [max] is 8 credits ($0.08). The price is fixed per image regardless of resolution within the model's supported range.

Input images are billed too

For editing and multi-reference work, the images you send in are also priced, not just the image that comes out. On FLUX.2, reference and input images are measured in megapixels and added to the cost of the generation. This is why the editing rate on FLUX.2 [pro] (from $0.045) is higher than its text-to-image rate (from $0.03). Exact per-megapixel input rates vary by model; the pricing calculator on the BFL site is the authoritative source for a specific resolution and input set.

What makes your bill higher than expected

Resolution

On FLUX.2, resolution is the main variable. Doubling the megapixels you ask for raises the per-image cost, and a 4MP render costs noticeably more than the one-megapixel floor in the table above. If you publish thumbnails or web images, generating at display resolution rather than maximum resolution keeps the bill down.

Editing costs more than generating

Editing and reference-guided generation bill for the input images on top of the output. On FLUX.2 [pro] that is the difference between $0.03 and $0.045 per image, and it climbs further when you pass several high-resolution reference images.

Batch requests do not get a discount

A batch request multiplies the base cost by the number of images. A batch of 4 with FLUX.2 [pro] costs 4 times the single-image price. Batching here is a convenience for issuing one request, not a way to pay less per image. This differs from text model APIs, where batch endpoints usually come with a discount.

Moving up a tier

The gap between tiers is large. FLUX.2 [max] at $0.07 is five times the cost of FLUX.2 [klein] 4B at $0.014 for the same one-megapixel image. Defaulting every call to the top model multiplies your bill across the whole workload.

How to reduce FLUX API costs

1. Match the model to the job

Model choice moves the bill the most. FLUX.2 [klein] 4B at $0.014 handles previews, drafts, and high-volume generation, while FLUX.2 [max] at $0.07 is for final, high-quality output. Generating drafts on klein and reserving max for approved finals can cut a workload's cost several times over compared with running everything on max.

2. Generate at the resolution you actually use

Because FLUX.2 is priced by megapixel, requesting only the resolution you need is a direct saving. Generate at display size and upscale separately the few images you keep, rather than rendering everything at maximum resolution up front.

3. Generate fewer candidates

Since batch gives no per-image discount, every extra candidate image is full price. Tightening prompts so you generate two options instead of eight, and only on the model tier the task needs, removes spend that batching will not.

4. Self-host the open-weight models

FLUX.2 [klein] and FLUX.1 [schnell] are available as open weights, so if you have GPU capacity you can run them without the per-image fee. This trades a usage fee for infrastructure and a license, and it carries restrictions: FLUX.2 [dev] is non-commercial unless you hold a commercial license, so check the licensing terms before shipping.

Can you use the FLUX API for free?

Puter.js: the User-Pays model

Puter.js is a JavaScript library that lets you add FLUX models to your app with no API key, no backend, and no bill to you as the developer. It works on the User-Pays model: each user of your app covers their own image generation through their Puter account, so your costs stay at zero no matter how many users you have.

<html>
<body>
  <script src="https://js.puter.com/v2/"></script>
  <script>
    puter.ai.txt2img(
      "A cat running like a human holding a silver fish in a market",
      { model: "black-forest-labs/flux-2-pro" }
    ).then(imageElement => {
      document.body.appendChild(imageElement);
    });
  </script>
</body>
</html>

We ran the workload we use across our pricing guides: 500 monthly users generating 20 images each, for 10,000 images a month. On FLUX.2 [klein] 4B at $0.014 per one-megapixel image, our calculation puts that at about $140 a month, growing linearly with your user base; on FLUX.2 [pro] the same 10,000 images run about $300 a month, and on FLUX.2 [max] about $700.

Through Puter.js the same app costs you $0 at any scale, because each user carries their own usage. At klein rates the absolute dollar saving is modest, but it still removes API key and billing management and stays free as you grow.

Self-hosting open weights

Running FLUX.2 [klein] or FLUX.1 [schnell] on your own hardware has no per-image fee, at the cost of GPU time and within the license terms.

Real-world cost examples

To make the numbers concrete, we priced three workloads at the rates above. All figures assume one-megapixel output unless noted; higher resolution raises the FLUX.2 totals.

An image generation feature in an app. We modeled 10,000 images a month and compared tiers, since model choice is the biggest single factor.

A monthly batch of marketing images. We calculated 1,000 finished images a month. On FLUX.2 [klein] 4B that is about $14, on FLUX.2 [pro] about $30, and on FLUX.2 [max] about $70. Rendering these at 2K or 4K instead of one megapixel raises each figure, so the calculator is worth checking for the exact resolution you ship.

An editing and personalization feature. We modeled 5,000 edits a month, where users adjust or restyle an existing image. On FLUX.1 Kontext [pro] at $0.04 that is about $200, and on FLUX.2 [pro] editing at $0.045 about $225. Reference images add to both totals because input is billed.

Complete FLUX API pricing table

FLUX.2 models (megapixel-based, "from" one megapixel)

FLUX.1 models (flat per image, 1 credit = $0.01)

FLUX.2 figures are the one-megapixel floor and rise with resolution; use the official calculator for an exact quote.

Self-hosting (open weights license)

Self-hosting tiers above Builder are quoted by BFL sales; see the licensing page.

Conclusion

FLUX.2 [max] starts at $0.07 per image and FLUX.2 [klein] 4B at $0.014, with the older FLUX.1 and Kontext models on flat credit pricing. To keep the bill down:

• Match the model to the job, klein for drafts and volume, max for finals.
• Generate at the resolution you actually use, since FLUX.2 is priced per megapixel.
• Generate fewer candidates, because batch is not discounted.
• Self-host open weights if you have the GPUs and the license fits.

Pricing verified on the BFL pricing docs.

Related

• Free, Unlimited FLUX AI API
• Best Image Generation APIs in 2026
• Free, Unlimited Image Generation API
• Ideogram API Pricing
• Stability AI API Pricing
• Free, Unlimited GPT Image API
• Free, Unlimited Imagen API
• Free, Unlimited Nano Banana API
• Free, Unlimited Stable Diffusion API
• Free, Unlimited Ideogram API
• OpenAI API Pricing
• Gemini API Pricing
• Claude API Pricing
• Free, Unlimited AI API