The GPT-4o native image generation model, which is far superior to DALL·E, has finally opened its API to the public.

The image generation API for GPT-4o, named gpt-image-1, has officially launched, offering two main features: first, generating images from scratch based on text prompts; and second, partially or fully modifying existing images using new prompts. These capabilities greatly enhance users’ flexibility in both creating and editing images. In addition, the API supports customization of image size and quality, as well as output with an alpha channel, further catering to diverse user needs. It is expected to have a positive impact on creatives and designers alike.