Generate image(s) via the gateway.

Returns images immediately for sync providers (Google, OpenAI) or an operation_id for async providers (Fal, Replicate).

Request body: model (str): Image model name prompt (str): Generation prompt number_of_images (int, optional): Number of images to generate (default: 1) input_images (list[str], optional): Input image URLs for edit mode <generation parameters> (optional): any parameter the model declares in its catalog schema (GET /v1/gateway/models/{model}/) — e.g. size / quality / background for the GPT Image family, rendering_speed / style for Ideogram, aspect_ratio / resolution elsewhere. Parameters the model does not declare are ignored; anything left out falls back to the model's declared default.