OpenAI's gpt-image-1 model brings Studio Ghibli-inspired image generation to developers
OpenAI has integrated the gpt-image-1 model, which powers the image generation feature in ChatGPT, into its API.
The move enables developers and enterprises to incorporate Studio Ghibli-inspired images into their applications.
The model enables high-quality image generation in a range of styles, adherence to custom guidelines and accurate rendering of text within images. OpenAI says this makes the model versatile for a wide range of practical applications.
OpenAI has revealed the pricing structure for the API. It’ll cost $5 per million tokens for text input, $10 per million for image input and $40 per million for generated images. That works out at approximately 2 cents, 7 cents and 19 cents per image for low, medium and high quality respectively.
The image generation feature has attracted more than 130 million users since its arrival in late March. Those users created 700 million images in the first week alone. The popularity has placed a significant strain on OpenAI’s resources.
Adobe, Canva, GoDaddy and Figma are among the companies already using gpt-image-1. Figma, for instance, is enabling users to generate and edit images on the fly within its platform.
OpenAI says all images generated through the API will include C2PA metadata that identifies them as content generated by AI. The company is also ensuring safety guardrails that are similar to those found within ChatGPT. Developers will be able to set the sensitivity of content moderation.
