DALL·E 3

DALL·E 3 is OpenAI's text to image system, introduced as a large step up in how closely a generated picture follows the words it was given. Earlier image models tended to skip or reorder parts of a description, which pushed users into learning prompt engineering. This system was built to read nuance and detail in ordinary sentences, so a short request or a long paragraph can both be rendered accurately. It was built natively on ChatGPT, which doubles as the prompting surface. You describe what you want in plain language, ChatGPT expands that into a detailed prompt, and you refine the result by asking for changes in conversation rather than rewriting the prompt yourself. Images created with it belong to the user, who needs no permission from OpenAI to reprint, sell or merchandise them. Safety work is visible in the product: requests naming a public figure are declined, as are requests for the style of a living artist, and creators can opt their images out of training for future image models. OpenAI does not price this model on its own. Image creation is bundled into the ChatGPT plans named Free, Go, Plus and Pro, whose amounts are rendered in the browser on the pricing page rather than written into the page itself. On the API price list this model no longer appears; current image generation is billed by token, quoted at $8.00 per million image input tokens, $2.00 per million cached image input tokens and $30.00 per million image output tokens, with text input at $5.00 per million. Treat this as a historical release. OpenAI's own page now carries a notice that the post introduced the system and points readers to ChatGPT Images 2.5 for the current experience.