promptstringrequired
Text description of the image you want to generate (max 1000 characters for dall-e-3, 4000 for dall-e-2)
modelstring
Image model to use — gpt-image-2 is the current flagship; dall-e-2/dall-e-3 are kept for existing steps built against them.
One ofgpt-image-2dall-e-3dall-e-2Default "gpt-image-2"
nnumber
Number of images to generate (1-10 for dall-e-2, 1 for dall-e-3)
Default 1
sizestring
Image size, in the chosen model's own format (e.g. "1024x1024", or "auto" for gpt-image-2). For dall-e-2: "256x256", "512x512", or "1024x1024". For dall-e-3: "1024x1024", "1792x1024", or "1024x1792". Left blank, the model uses its own default.
qualitystring
Quality setting — dall-e-3 only ("standard" or "hd"); not applied for other models.
One ofstandardhdDefault "standard"
seednumber
Seed for image generation (for reproducibility)
stylestring
Style for dall-e-3 only ("vivid" or "natural")
One ofvividnaturalDefault "vivid"
userstring
Unique identifier for the end-user
