Black Forest Labs: FLUX.2 [klein] 9B Base
flux-2-klein-9b-base- Price
- From $0.00416
- Modalities
- Image
About FLUX.2 [klein] 9B Base
FLUX.2 [klein] 9B Base is the undistilled foundation model of Black Forest Labs' klein family, giving the full 9-billion-parameter capacity for image generation and editing. It is meant for fine-tuning, customization, and post-training, where flexibility and control matter more than the few-step speed of the distilled klein 9B.
Where it works well
- It is the undistilled foundation of the klein family, so training keeps the maximum signal from the base weights.
- Black Forest Labs positions it for fine-tuning, customization, and post-training rather than only for inference.
- The 9 billion parameters give more capacity than the 4B Base for detailed scenes and edits.
- It handles text-to-image and image editing in one model, including multi-reference input.
When to choose another model
- Because it is undistilled it samples more slowly than the 9B distilled model, which makes it a poor choice for real-time use.
- Black Forest Labs says rendered text may be distorted and that prompt-following depends on phrasing style.
- The weights follow the FLUX Non-Commercial License, unlike the Apache-licensed 4B variants.
Getting started
Create API key
Create a key in Console, then use it with every model on the platform.
Send your first request
Copy the example for your language and run it against the endpoint.
Image editingText to imageTokenLab endpointPOST/v1/images/generationscurl -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-klein-9b-base", "prompt": "A minimalist product photo of matte black headphones on a soft blue background.", "image_url": "https://example.com/source.png" }'
Parameters
Image editing
| Parameter | Values and limits | Default | Notes |
|---|---|---|---|
cfg_scale | ≥ 1, ≤ 20 | Not provided | Use multiples of 0.01. |
guidance_scale | ≥ 1, ≤ 20 | Not provided | Use multiples of 0.01. |
height | ≥ 128, ≤ 2048 | 1024 | Use an integer. Use multiples of 16. |
image | Not provided | Not provided | Alias of the combined images input. See input_limits.images for the total input image count. |
image_url | Not provided | Not provided | Alias of the combined images input. See input_limits.images for the total input image count. |
image_urls | Not provided | Not provided | Alias of the combined images input. See input_limits.images for the total input image count. |
images | Inputs: ≥ 1, ≤ 4 | Not provided | Limits apply when this input is supplied. Image aliases are combined into one input image list; the limit applies to the combined list. |
n | 1 | 1 | Use an integer. |
negative_prompt | Not provided | Not provided | Minimum length: 1 characters. Maximum length: 10000 characters. |
operationRequired | Not provided | Not provided | Not provided |
output_format | jpeg, png, webp | Not provided | Default encoding is shared by these equivalent values: jpg, jpeg. |
prompt | Not provided | Not provided | Minimum length: 1 characters. Maximum length: 10000 characters. |
seed | ≥ 0, ≤ 9007199254740991 | Not provided | Use an integer. |
size | Not provided | 1024x1024 | Use WIDTHxHEIGHT in pixels. The width and height constraints also apply to size. |
steps | ≥ 1, ≤ 50 | 28 | Use an integer. |
width | ≥ 128, ≤ 2048 | 1024 | Use an integer. Use multiples of 16. |
Text to image
| Parameter | Values and limits | Default | Notes |
|---|---|---|---|
cfg_scale | ≥ 1, ≤ 20 | Not provided | Use multiples of 0.01. |
guidance_scale | ≥ 1, ≤ 20 | Not provided | Use multiples of 0.01. |
height | ≥ 128, ≤ 2048 | 1024 | Use an integer. Use multiples of 16. |
images | Not provided | Not provided | Text-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. |
n | 1 | 1 | Use an integer. |
negative_prompt | Not provided | Not provided | Minimum length: 1 characters. Maximum length: 10000 characters. |
operationRequired | Not provided | Not provided | Not provided |
output_format | jpeg, png, webp | Not provided | Default encoding is shared by these equivalent values: jpg, jpeg. |
prompt | Not provided | Not provided | Minimum length: 1 characters. Maximum length: 10000 characters. |
seed | ≥ 0, ≤ 9007199254740991 | Not provided | Use an integer. |
size | Not provided | 1024x1024 | Use WIDTHxHEIGHT in pixels. The width and height constraints also apply to size. |
steps | ≥ 1, ≤ 50 | 28 | Use an integer. |
width | ≥ 128, ≤ 2048 | 1024 | Use 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
- $0.00416
- Verified price
- —
- Discount
- —
Text to image
per request- Official price
- $0.00416
- Verified price
- —
- Discount
- —
| Official priceper request | Verified priceper request | Discount | |
|---|---|---|---|
| Image editing | $0.00416 | — | — |
| Text to image | $0.00416 | — | — |
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
Data is based on aggregate user requests, excluding status checks.
Open in Console
Open FLUX.2 [klein] 9B Base in Console with a prompt ready to edit or send.
Help me create an image with flux-2-klein-9b-base using image-edit at /v1/images/edits. Show the result, cost, and any limits I should know.
Use cases
Best for- Text to Image
Custom model training
Start a LoRA or full fine-tune for a character, product line, or art direction from a checkpoint that has kept its full capacity.
Large-scale variation
Draw many distinct images from one prompt when a team needs options rather than one fast answer.
Editing pipelines
Combine a base image with one or more references and an instruction, then compare with the distilled model on the same inputs.
Evaluation baselines
Measure how much quality and control the distilled klein 9B gives up by running both on one prompt set.
Prompt examples
A 1970s travel poster for a lunar colony, screen-print texture, three-color palette, no text.
Same chair as in the reference image, now in forest green velvet, photographed in a gallery with raking light.
Portrait of an elderly fisherman, overcast harbor, 85mm look, natural skin texture.
FAQ
What is FLUX.2 [klein] 9B Base?
It is the undistilled foundation model of the klein family from Black Forest Labs. It keeps the full 9-billion-parameter capacity and is optimized for fine-tuning, post-training, and workflows that want flexibility and control.
How does it differ from FLUX.2 [klein] 9B?
The plain 9B is step-distilled to four inference steps for sub-second results. The Base keeps the undistilled weights, so it is slower to sample but better as a training starting point and gives more varied results.
Should I pick 9B Base or 4B Base?
Pick 9B Base for more capacity on detailed scenes and edits, and when its non-commercial license suits you. Pick 4B Base for lower memory use and Apache 2.0 licensing.
Does it edit images as well as generate them?
Yes. The catalog lists both text-to-image and image-edit operations, and the klein family supports multi-reference editing in a single model, so one checkpoint serves both kinds of request.
Is it a good default for production image generation?
Usually not. For fast, repeated calls the distilled klein 9B or a larger FLUX.2 model is the better fit. Choose the Base when you want output variety or plan to adapt the model yourself.
How much does FLUX.2 [klein] 9B Base cost?
On TokenLab, FLUX.2 [klein] 9B Base costs $0.00416 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 [klein] 9B Base use?
Use https://api.tokenlab.sh/v1/images/edits for FLUX.2 [klein] 9B Base. The request example below shows the matching code shape.
Which operations does FLUX.2 [klein] 9B Base support?
FLUX.2 [klein] 9B Base supports Image editing, Text to image. Select an operation above to see its endpoint and request example.
Compare FLUX.2 [klein] 9B Base
Sources
Reviewed Oct 2, 2026