GPT Image is OpenAI's family of image generation models, offered through the OpenAI API, an interface that other programs call over the internet. GPT Image 2 (model gpt-image-2, snapshot of 21 April 2026) generates and edits images from text and image input, supports flexible image sizes and accepts high-fidelity image inputs. The image generation guide now names two newer models: gpt-image-2.5-sunburst, for edits that need precision, and gpt-image-2.5-flare, for fast everyday generation.
Developers reach the models in two ways. The Image API has a Generations endpoint, which creates an image from a text prompt, and an Edits endpoint, which changes part or all of an existing image. The Responses API offers image generation as a built-in tool inside a conversation, which allows multi-turn editing and file IDs as image inputs. Both set quality, size, file format and compression. An organization may have to complete API Organization Verification before it can use GPT Image models.
Use is billed per token, a small unit of text or image data, with separate prices for text input, image input and image output; GPT Image 2 is not available on the free API tier.