FLUX 3 Image Klein API and Pricing, Explained Through a Real Pipeline
FLUX 3 Image Klein is the open-weights member of Black Forest Labs' FLUX 3 family: the same five resolution tiers and up to 10 reference images as FLUX 3 Image, at under a cent per image for every tier below 4K. Here is the rate card, what we measured on the day it landed, where it fits next to its bigger sibling, and how to reach it through a published API endpoint.
FLUX 3 Image Klein is the open-weights image model in Black Forest Labs' FLUX 3 family. It reached OpenRouter on October 16, 2026, as black-forest-labs/flux-3-image-klein, and we added it to the Image Generation node, to chat, and to the model pages the same morning. It does the two jobs its bigger sibling does: it generates from a prompt alone, and it edits or composes from up to 10 reference images.
What is different is the price. Klein lists at less than a cent per image at every tier up to 2K, and its 4K tier costs less than FLUX 3 Image charges for a 768 draft. This post covers the rate card, what we measured when we onboarded the model, how it compares with FLUX 3 Image, and how to call it as a published API endpoint.
The spec sheet
| Capability | FLUX 3 Image Klein |
|---|---|
| Resolution tiers | 768, 1K, 1.5K, 2K, 4K |
| Aspect ratios | 15, from 21:9 through 9:21, plus auto |
| Input | Text and image |
| Reference images | Up to 10 per request |
| Images per request | 1 |
| Weights | Open |
The tiers, the aspect ratio list, and the reference ceiling are identical to FLUX 3 Image, so a pipeline built around one can swap to the other by changing the model on the node and nothing else. The full capability list is on the FLUX 3 Image Klein model page.
What FLUX 3 Image Klein costs per image
These are the listed prices for the model as of October 2026, with FLUX 3 Image beside them for scale:
| Resolution | FLUX 3 Image Klein | FLUX 3 Image |
|---|---|---|
| 768 | $0.008 | $0.041 |
| 1K | $0.009 | $0.048 |
| 1.5K | $0.011 | $0.07 |
| 2K | $0.014 | $0.10 |
| 4K | $0.036 | $0.607 |
Three things in that table are worth reading closely.
Everything below 4K costs about a cent. The spread from 768 to 2K is six tenths of a cent, from $0.008 to $0.014. There is no reason to draft at 768 to save money; the saving between 768 and 2K on a hundred images is sixty cents.
4K is cheap enough to be a default. Klein's 4K tier lists at $0.036, which is less than FLUX 3 Image charges at 768. The 4K warning in our FLUX 3 Image post does not apply here. Twenty 4K renders list at $0.72 on Klein and $12.14 on FLUX 3 Image.
The gap widens as resolution rises. Klein is about five times cheaper at 768 and 1K, about seven times cheaper at 2K, and about seventeen times cheaper at 4K.
References cost nothing extra on either model, with ten attached or with none.
On the node, leaving Resolution on Model default renders a 1024 by 1024 square at the 1K price. Pick a tier explicitly when you want anything else.
What we measured on launch day
We probed the model three times on October 16, 2026, the day it landed, through the same path the Image Generation node uses. The billed cost matched the listed price in every case:
| Probe | Setting | Billed | Wall time |
|---|---|---|---|
| Text to image | 768, 1:1 | $0.008 | about 7 seconds |
| Text to image | Resolution unset, 1:1 | $0.009 | about 5 seconds |
| Edit with one reference | 768, 1:1 | $0.008 | about 10 seconds |
The unset run confirmed the default: a 1024 by 1024 square at the 1K price. The reference run used a product still of a folded navy linen shirt and asked for the same shirt on a white marble counter in a bright studio. The result kept the shirt, its color, fold, collar and buttons, and placed it on marble under window light. That is the edit most product teams want, and it cost eight tenths of a cent.
Speed is the other half of the story. FLUX 3 Image came back in 23 to 49 seconds across our probes on October 4. Klein came back in 5 to 10. Three renders from one day is a small sample, not a benchmark, but the difference was consistent enough that Klein went straight onto the chat allowlist, where a model has to answer quickly to be useful.
Which model to use for what
With two FLUX 3 models on the node, the question is how to split work between them. The rate card suggests a simple pattern:
- Draft on Klein, at whatever tier you like. A hundred variations of a prompt at 1K list at $0.90. Explore composition, lighting, and layout here.
- Batch on Klein. Storyboards, thumbnail candidates, carousel frames, and product sets are volume jobs, and a model under a cent per frame that answers in seconds is built for them.
- Finish on FLUX 3 Image when a frame earns it. Run the chosen prompt and references through the bigger model on a second node. If it looks better for that subject, pay for it once rather than a hundred times.
We are not going to tell you which model looks better, because that depends on your subject. Put the same prompt and references through both on two nodes and compare. Seedream 5.0 Flash is worth a third node if you want a low-cost model from a different lab in the test.
Reference images and consistent sets
With a prompt alone Klein generates. With a prompt plus references it edits or composes, on the same node.
The ceiling of 10 references is what makes it useful for sets rather than single images. Attach a character, a product, and a setting, then ask for the next frame in a series, and each result carries those references forward. Set the aspect ratio to auto and the output follows the shape of the reference. Set it to 9:16 and the same references render as vertical frames for Stories and Shorts.
Because every frame costs under a cent, generate several candidates per shot, keep the one that matches best, and move on. The cleaner the reference still, the less the result drifts.
The endpoint detail that trips up integrations
If you call Klein through OpenRouter yourself, note that it runs on the images endpoint only. When we probed it, the chat completions route refused the request, the same as FLUX 3 Image. An integration written for chat-completions image output needs a second code path for this model.
Inside a Treza pipeline you never see this. The Image Generation node knows which models require the images endpoint and routes the request there. You pick FLUX 3 Image Klein in the model menu and the node handles the rest.
What you actually pay on Treza
FLUX 3 Image Klein through a Treza pipeline is metered from a prepaid credit balance. One credit equals $0.01, packs start at $5, credits never expire, and there is no subscription. A failed generation charges nothing, and that includes a prompt the provider refuses.
The credit estimate on the node uses the listed price for the tier you pick, so the number you see before a run is the number you pay for a successful render. At Klein's prices, most runs round to a single credit.
Calling FLUX 3 Image Klein as an API
Direct access means provider credentials, the images endpoint request shape, error handling, and retry logic before you have built anything for your product.
Through a Treza pipeline, Klein is a node on a canvas. Place the Image Generation node, pick the model, set resolution and aspect ratio, wire references into its reference input, and publish the graph as a single versioned endpoint. Your app calls the typed /invoke route with JSON in and JSON out, or points an existing OpenAI SDK at the OpenAI-compatible route. Swap the node to FLUX 3 Image later and the endpoint your product calls stays the same.
The same pipeline is reachable from an agent. Treza's hosted MCP server exposes pipelines as tools that Claude and other MCP clients call directly, so an agent can generate the image and hand back the URL as one step of a larger task.
Putting it in a pipeline instead of a single call
A single image is rarely the deliverable. The more useful pattern is Klein as the first node of several.
Compose a key frame from your references, then pass it to a video node as the first frame. A video model does not need a 4K source, so a 1K or 2K still is the sensible input, and the post on setting a first-frame image for AI video generation covers that handoff. The image to video generator page has that pairing set up, and the AI thumbnail generator page covers the other end of the same graph, where a thumbnail is built from the frames of the finished video.
A fuller AI video pipeline puts a script step ahead of the image and narration, captions, or publishing behind the video, then reruns the whole graph on a schedule or from an API call. The AI image generator page has the image side set up, the AI video generator page puts a video model behind a single generate action, and how to build an AI image generation pipeline walks through image graphs in more depth.
Frequently Asked Questions
How much does FLUX 3 Image Klein cost per image?
The listed prices as of October 2026 are $0.008 at 768, $0.009 at 1K, $0.011 at 1.5K, $0.014 at 2K, and $0.036 at 4K. The price depends only on the resolution tier. Aspect ratio and the number of reference images do not change it.
How does FLUX 3 Image Klein compare with FLUX 3 Image?
Both take up to 10 references, render at the same five tiers, and offer the same fifteen aspect ratios plus auto. Klein is the open-weights model and lists at about five times less at 768 and 1K, seven times less at 2K, and seventeen times less at 4K. Only a side-by-side run of your own prompt can say which looks better for your subject.
Do reference images cost extra with FLUX 3 Image Klein?
No. The model accepts up to 10 reference images per request and lists no charge for them. When we probed it, an edit with one reference billed the same $0.008 as a plain text prompt at the same tier.
What resolution does FLUX 3 Image Klein render by default?
On the Image Generation node, leaving Resolution on Model default renders a 1024 by 1024 square at the 1K price. Choose 768, 1.5K, 2K, or 4K on the node when you want a different size.
How fast is FLUX 3 Image Klein?
In our three probes on October 16, 2026, images came back in about 5 to 10 seconds. That is a small sample from one day, so treat it as a rough expectation rather than a benchmark.
Does FLUX 3 Image Klein work through the chat completions API?
Not on OpenRouter as of October 2026. The model runs on the images endpoint, and the chat completions route refused it when we tested. On Treza the Image Generation node routes the request to the right endpoint automatically.
Is FLUX 3 Image Klein free?
Building a pipeline on the canvas costs nothing, but generation spends real provider money and is charged from a prepaid credit balance. There is no subscription, packs start at $5, credits never expire, and a failed generation charges nothing.
