OpenAI is opening up access to a new image generation system via an API, allowing developers to integrate it into their apps and services, TechCrunch reports.
The updated generator, which became available to ChatGPT users in late March, has gained popularity for its ability to create realistic Ghibli-style images. In the first week of operation, over 130 million ChatGPT users created over 700 million images.
The new generator is based on the multimodal artificial intelligence model gpt-image-1, which supports different image styles, adapts to user preferences, has extensive knowledge of the world, and displays text.
Developers will be able to generate multiple images at once, control their quality and speed of creation. However, the model has protective mechanisms that limit the creation of content that contradicts the company's policy. OpenAI emphasizes that all images created using their model contain watermarks confirming their origin from artificial intelligence.
The cost of generation depends on the quality of the image: from 2 to 19 cents per square image. Companies already using gpt-image-1 include Adobe, Wix, Instacart, Canva, and Figma.