VidMachine

VidMachine

Black Forest Labs

Flux 2 Klein

Black Forest Labs · AI image
FLUX.2 klein 4B is a compact, open-source image generation and editing model from Black Forest Labs, distilled from the larger FLUX.2 base model. At 4 billion parameters it is tuned for speed—sub-second inference in many cases—while keeping the lighting, materials, and composition quality inherited from its larger teacher model.
It is released under Apache 2.0, so commercial use carries no licensing fees. For teams that need reliable photorealistic output without the cost and latency of flagship models, klein is the efficiency-focused entry in the FLUX.2 family: klein for real-time use, pro for production scale, flex for fine-grained control, and max for highest quality with grounding.
On VidMachine, Flux 2 Klein is priced at 1 credit per image, making it the cheapest way to generate and edit images—ideal for iterating on ideas, bulk scene drafts, and high-volume shorts production.

Key features and benefits

Speed and efficiency

The 4B model uses latent flow matching with a 4-step distilled sampler, delivering sub-second inference on Replicate. That speed makes it practical for near real-time applications, rapid iteration, and production pipelines where latency matters.

Text rendering

Unlike many small image models, FLUX.2 klein produces readable typography in complex layouts, infographics, and UI mockups. Text-heavy prompts that would garble on other compact models often come out legible on the first attempt.

Spatial reasoning

The model understands how lighting works in the real world: shadows fall where they should, reflections behave correctly, and objects keep proper perspective relationships. This reduces the tell-tale "AI look" that makes generated images feel off.

High-resolution editing

You can edit images at resolutions up to 4 megapixels while preserving detail and coherence. The model maintains geometry and texture during edits rather than hallucinating new details.

Multi-reference support

Provide multiple reference images to maintain consistent characters, products, or styles across generations—useful for serial content, product variations that feel like one brand, or the same person appearing across scenes.

Technical specifications

Parameters4 billion (distilled from FLUX.2 base)
OutputUp to 4 megapixels
Reference imagesUp to 5
LicenseApache 2.0 (free for commercial use)
ArchitectureLatent flow matching; Mistral 3 vision-language encoder + rectified flow transformer
AccessPlayground, API, open weights on Hugging Face

Use cases and applications

FLUX.2 klein fits workflows where you need reliable output without high-end hardware or flagship pricing: product visualization with consistent branding across variants, UI mockups and interface prototypes with readable text, marketing materials that must match exact brand colors (hex codes work in prompts), and character consistency across storyboards or campaigns.
On VidMachine it is a strong default for drafting scenes quickly—generate several options at 1 credit each, pick the best composition, then refine with reference images or re-run with a different model when a hero frame needs maximum fidelity.

Why this model

Flux 2 Klein is the value pick in the FLUX.2 line: most of the composition quality, text rendering, and editing capability of the larger variants at a fraction of the cost and latency. Choose it when iteration speed and price per image matter more than absolute top-tier quality.
For briefs that demand maximum resolution, more reference images, or the last few percent of fidelity, compare Flux 2 Pro, Seedream, or Nano Banana 2 on your criteria.

How VidMachine uses it

VidMachine runs Flux 2 Klein through Replicate at 1 megapixel output with your project's aspect ratio, so 9:16 shorts and 16:9 landscape frames come out ready for the timeline.
It accepts up to 5 reference images for character, product, and style consistency. Each generated image costs 1 credit, regardless of how many references you attach.

What you should know

How much does Flux 2 Klein cost on VidMachine?
1 credit per generated image. Reference images are included—attaching up to 5 references does not change the price.
What is the difference between FLUX.2 variants?
FLUX.2 klein is the compact, fast variant for real-time use; pro targets production scale; flex offers fine-grained control; max delivers the highest quality with grounding. Klein is open source under Apache 2.0, so it is also free to self-host.
How many reference images can Flux 2 Klein use?
Up to 5 input images for image-to-image generation and editing, useful for character consistency and product variations.
Can I use Flux 2 Klein commercially?
Yes. FLUX.2 klein is released under the Apache 2.0 license, so you can use it commercially, modify it, and build on top of it without licensing fees.