Edit and transform existing images using Kling O3 with natural language instructions. Supports up to 10 reference images, 1K/2K/4K resolutions, and up to 9 outputs per request.
Edit and transform existing images using Kling O3 with natural language instructions. Supports up to 10 reference images, 1K/2K/4K resolutions, and up to 9 outputs per request.
About this model
Kling O3 Image Edit transforms existing images using natural-language instructions, with up to 10 reference images per request, selectable 1K / 2K / 4K resolution, multiple aspect ratios (including auto to follow the reference), and up to 9 outputs per request.
Cost analysis
| Provider | Cost | Notes |
|---|---|---|
| muapiapp | $0.027 – $0.055 per image | 1K / 2K: $0.027 per image. 4K: $0.055 per image. Scales linearly with the number of images requested, regardless of how many references you upload. |
| Fal.ai | $0.028 – $0.056 per image | 1K / 2K: $0.028 per image. 4K: $0.056 per image. |
| Replicate | Not available | Kling O3 Image Edit is not available on Replicate. |
1K / 2K: $0.027 per image. 4K: $0.055 per image. Scales linearly with the number of images requested, regardless of how many references you upload.
1K / 2K: $0.028 per image. 4K: $0.056 per image.
Kling O3 Image Edit is not available on Replicate.
** Competitor pricing is estimated based on similar model architectures and usage tiers.
Configuration schema
| Parameter | Type | Description | Default |
|---|---|---|---|
| Prompt | string | Text instructions describing the desired transformation. Maximum 2,000 characters. | Preserve the original composition, people placements, facial identities, DJ setup, and rooftop party mood from the source image. Transform the entire scene into a tiny miniature rooftop party built on top of a moving RC toy truck driving across a messy apartment floor. Surround the tiny party with giant everyday household objects like shoes, cables, snack packets, books, and soda cans towering like skyscrapers. Keep the same party energy and realistic interactions while introducing playful scale contrast, macro photography depth, cinematic lighting, and highly detailed miniature-world realism. |
| Image URLs | array | Upload or provide reference images to transform. Up to 10 images supported. | https://d3adwkbyhxyrtq.cloudfront.net/webassets/videomodels/kling-o3-image-edit-in.jpg |
| Aspect Ratio | Enum (9 options) | Output image aspect ratio. Use 'auto' to follow the reference image. | auto |
| Resolution | Enum (3 options) | Output image resolution. | 1K |
| Number of Images | int | How many images to generate per request. | 1 |
Text instructions describing the desired transformation. Maximum 2,000 characters.
Preserve the original composition, people placements, facial identities, DJ setup, and rooftop party mood from the source image. Transform the entire scene into a tiny miniature rooftop party built on top of a moving RC toy truck driving across a messy apartment floor. Surround the tiny party with giant everyday household objects like shoes, cables, snack packets, books, and soda cans towering like skyscrapers. Keep the same party energy and realistic interactions while introducing playful scale contrast, macro photography depth, cinematic lighting, and highly detailed miniature-world realism.Upload or provide reference images to transform. Up to 10 images supported.
https://d3adwkbyhxyrtq.cloudfront.net/webassets/videomodels/kling-o3-image-edit-in.jpgOutput image aspect ratio. Use 'auto' to follow the reference image.
autoOutput image resolution.
1KHow many images to generate per request.
1Developer documentation
Upload your images: Provide up to 10 reference images via the images_list field.
Write your prompt: Describe the transformation you want — target style, composition changes, elements to keep or replace. Up to 2,000 characters are supported.
Pick aspect ratio and resolution: Use auto to follow the reference, or pick a specific aspect ratio. Choose 1K, 2K, or 4K resolution.
Choose how many images: Set num_images between 1 and 9. The total cost scales linearly with the number of outputs.
Submit: Click Generate. The model applies your instructions to the references and returns the requested number of images.
Frequently asked
Between 1 and 10 images via the `images_list` field. The model will use them as visual references when generating the output.
It lets the model pick the output aspect ratio based on your reference images, which is usually the right default for edit jobs.
Between 1 and 9 outputs per request. You're billed per generated image.