Tasks and inputs
Krea 3 is listed for Image-text-to-image, Image-to-image, Text-to-image. It accepts text, image and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
Krea 3 is Krea's in-house image model for text-to-image and instruction-based edits at 1K or native 2K, priced from $0.03 per image.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: Indicative: $0.03 per 1K output image, $0.06 per 2K. A $0.07 per input reference image listing on Cloudflare AI Gateway is unconfirmed. Confirm on your account.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Krea 3 is the model behind the Krea app, and it marks a shift for the company: no Krea 3 weights have been released, unlike Krea 2. What you get instead is an in-house model, launched as a beta on 15 October 2026, that handles both text-to-image generation and instruction-based editing in one place.
Editing is the headline. Krea says most 2K edits come back in under eight seconds end to end. That speed changes the workflow: you iterate on an image the way you would in an editor, not the way you queue a render.
Krea 3 accepts an image to edit plus several reference images. The distinctive part is that each reference is described by what it contributes. One image supplies the lighting, another the material, another the pose. That is more precise than dropping in a mood board and hoping the model infers intent.
The API also exposes:
| Output | Price |
|---|---|
| 1K image | $0.03 |
| 2K image | $0.06 |
A registry also lists $0.07 per input reference image through Cloudflare AI Gateway, which we could not confirm from Krea's own pages. If you use many references per job, check that charge on your account, because it could exceed the output cost.
Beyond Krea's own app and API, the model is distributed through partners including fal, Replicate, Together AI and OpenRouter. Prices can differ by partner.
Some coverage mentions output up to 4K for print. Krea's documentation describes 1K and native 2K only, so plan on 2K. And while most coverage uses 15 October as the launch date, some places it on 12 October.
Generation demos flatter every model. Test the edits. Take ten real images from your work, write one precise instruction for each ("swap the jacket to navy, keep the stitching"), and add a reference image for half of them. Score each result on whether the instruction was followed and whether untouched regions stayed untouched. Repeat with a regional edit. Then run the same set through a rival such as Nano Banana 2 and compare cost per accepted image, retries included.
More options are under image-to-image models and text-to-image models. Krea's open releases are on the Krea provider page.
Compare the model on the work you need to complete.
Krea 3 is listed for Image-text-to-image, Image-to-image, Text-to-image. It accepts text, image and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
Choose the output size and quality your design needs. Include drafts and revisions when budgeting an accepted image.
Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
recraft
Recraft V4.1 Flash V2 is Recraft's fastest, cheapest image model: 1K raster output in about 1.1 s for $0.007, quality close to Recraft V4.1.
MiniMax
MiniMax H3 Image is an open-weight 33.1B image generation and editing model producing native ~16MP output in six steps, with up to 16 reference images.
MiniMax
MiniMax H3 Image Highspeed is the faster hosted tier of MiniMax's H3 Image generator and editor, at double the image-token price of the standard tier.
OpenAI
GPT-Image-2.6 Flare is OpenAI's fast image generation and editing model for interactive iteration, with SVG, animated GIF and markup-guided edits.
OpenAI
GPT-Image-2.6 Sunburst is OpenAI's detail-first image generation and editing model, adding SVG, animated GIF output and markup-guided edits.
Alibaba
Qwen-Image-2.1-Turbo creates and edits images, including reference-guided and transparency workflows; hosted billing is separate from local deployment.