interface ImageGenerateParams {
    model?: null | string & {} | ImageModel;
    n?: null | number;
    prompt: string;
    quality?: "standard" | "hd";
    response_format?: null | "url" | "b64_json";
    size?:
        | null
        | "256x256"
        | "512x512"
        | "1024x1024"
        | "1792x1024"
        | "1024x1792";
    style?: null | "natural" | "vivid";
    user?: string;
}

Properties

model?: null | string & {} | ImageModel

The model to use for image generation.

n?: null | number

The number of images to generate. Must be between 1 and 10. For dall-e-3, only n=1 is supported.

prompt: string

A text description of the desired image(s). The maximum length is 1000 characters for dall-e-2 and 4000 characters for dall-e-3.

quality?: "standard" | "hd"

The quality of the image that will be generated. hd creates images with finer details and greater consistency across the image. This param is only supported for dall-e-3.

response_format?: null | "url" | "b64_json"

The format in which the generated images are returned. Must be one of url or b64_json. URLs are only valid for 60 minutes after the image has been generated.

size?:
    | null
    | "256x256"
    | "512x512"
    | "1024x1024"
    | "1792x1024"
    | "1024x1792"

The size of the generated images. Must be one of 256x256, 512x512, or 1024x1024 for dall-e-2. Must be one of 1024x1024, 1792x1024, or 1024x1792 for dall-e-3 models.

style?: null | "natural" | "vivid"

The style of the generated images. Must be one of vivid or natural. Vivid causes the model to lean towards generating hyper-real and dramatic images. Natural causes the model to produce more natural, less hyper-real looking images. This param is only supported for dall-e-3.

user?: string

A unique identifier representing your end-user, which can help OpenAI to monitor and detect abuse. Learn more.