Find the Framing
Swap in your subject and keep the structure.
A street photograph of a cyclist passing a pale blue wall at sunrise. Wide composition, long shadows, natural colors, sharp subject.
Explore visual directions, fast composition drafts, and color variations with FLUX.2 [klein]. Supports up to four input references with 9B speed.
Start with an idea
Describe what you want to create. Add a reference if you need one, then adjust the options below.
Explore visual directions, fast composition drafts, and color variations with FLUX.2 [klein]. Supports up to four input references with 9B speed.
Find a visual direction, then make the prompt your own.

Early variations in composition, color and mood.
Example prompt
A bright, surreal scene on a vast white salt flat under a clear blue sky, featuring two fluffy white alpacas standing in the foreground. Behind them sits a sleek white supercar with its scissor doors open, creating a dramatic futuristic silhouette. The sunlight is strong and crisp, casting sharp shadows on the ground and giving the image a clean, high-contrast cinematic look. Minimalist composition, playful luxury vibe, ultra sharp detail, wide-angle perspective, high resolution.
Swap in your subject and keep the structure.
A street photograph of a cyclist passing a pale blue wall at sunrise. Wide composition, long shadows, natural colors, sharp subject.
Upload your images before generating.
Keep the subject and composition in @image1. Change warm sunset light to cool overcast daylight. Preserve camera angle and all objects.
Credit costs depend on the selected settings and reference inputs.
Each cell is the total for one image, not an additional reference charge. Reference counts within the same band have the same price.
| Settings | Generate · no references | Edit · 1 reference images | Edit · 2–4 reference images |
|---|---|---|---|
| Standard output | 10 credits | 10 credits | 25 credits |
| Settings | Generate · no references | Edit · 1 reference images | Edit · 2–4 reference images |
|---|---|---|---|
| Large output / Auto edit | 10 credits | 10 credits | 33 credits |
Standard covers outputs up to 2,097,152 pixels, including all preset ratios. Large covers larger supported outputs up to 4,194,304 pixels. Auto-sized edits use the Large tier.
Once your direction is clear, compare Pro to develop a finished image.
Compare the settings available for your next image.
| Settings | Flux.2 KleinCurrent model | Flux.2 Pro | Flux.2 Max | Flux.2 Flex | Flux 3 | Seedream 5.0 Lite | Seedream 5.0 Pro | Nano Banana 2 | Nano Banana 2.1 | MiniMax H3 Image | GPT Image 2.6 | GPT Image 2.5 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Resolution tiers | Standard | Standard | Standard | Standard | 1K | 2K / 3K / 4K | 1K / 2K | 0.5K / 1K / 2K / 4K | 1K / 2K / 4K | Source size when editing | Source size when editing | Source size when editing |
| Quality levels | Standard | Standard | Standard | Standard | Standard | Standard | Standard | Standard | Standard | Low / Medium / High | Low / Medium / High | Low / Medium / High |
| Reference images | Up to 4 | Up to 4 | Up to 4 | Up to 4 | Up to 4 | Up to 4 | Up to 4 | Up to 4 | Up to 4 | Up to 4 | Up to 4 | Up to 4 |
| Images per generation | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| Base credits | 10 credits | 17 credits | 38 credits | 38 credits | 20 credits | 14 credits | 54 credits | 32 credits | 24 credits | 30 credits | 30 credits | 16 credits |
| Base credits by tier | Standard output: 10 | Standard output: 17 | Standard output: 38 | Standard output: 38 | 1K: 20 | 2K: 14 / 3K: 14 / 4K: 14 | 1K: 27 / 2K: 54 | 0.5K: 24 / 1K: 32 / 2K: 48 / 4K: 63 | 1K: 24 / 2K: 32 / 4K: 63 | Low quality: 25 / Medium quality: 30 / High quality: 60 | Low quality: 25 / Medium quality: 30 / High quality: 60 | Low quality: 8 / Medium quality: 16 / High quality: 32 |
Credit costs depend on the selected settings and reference inputs.
Inputs, editing, and credits—before you start.
Yes. The paid 9B variant supports up to four references here; the free model supports single-image editing.
Flux.2 Klein is built for fast drafts and visual variations. You can generate images from text briefs, edit uploaded photos with natural language, or guide composition with up to 4 reference images. All creations can also be used as starting frames for AI video generation.
FluxContext supports up to 4 reference images for Flux.2 Klein. You can assign roles (Subject, Background, Style) and upload JPG, PNG, or WebP files up to 4MB each.
Available settings: Standard. Choose your preferred aspect ratio, framing, and quality tier. Adding references helps guide identity, background, and style.
Generation starts from 10 credits for standard aspect ratios. Pricing adjusts transparently based on your resolution, quality tier, and reference inputs. See the pricing information on this page for model credit costs.
Yes. Click Edit on any delivered image to continue refining with natural language while preserving earlier versions in your history. You can also click Generate video to turn it into an animated clip.
Yes. Flux.2 Klein is a premium production model that requires signing in and credits. New accounts receive 10 free credits to test models upon sign-up, or you can try our free FLUX.2 model without an account.
AI image editing interprets your prompt within the visual context rather than locking exact pixels. To maximize detail preservation, clearly specify which elements must remain untouched and describe only the changes you need.
Describe your idea and start creating.