Openai allows the generator to developed for developers

On Wednesday, Openai brought technology behind the new and improved image generation feature in its Chatgpt to its API, allowing developers to merge them into their applications and services.

The new photo generator from Openai, which It was launched for most Chatgpt users in late MarchHer viral gold for her The ability to create realistic images in GHibli And “and”Artificial intelligence work numbers“It was a mixed blessing for Openai, which led to millions of new Chatgpt subscriptions with a great connection to the company’s ability.

In the Openai Application interface, the image generation capacity is run by the AI ​​model called “GPT-ISMAGE-1”. GPT-amge-1, a multimedia model, can create images across different patterns, follow the designated guidelines, lift global knowledge, and a text of display.

Developers can create multiple images simultaneously using GPT-amge-1, and control the quality-and the speed.

According to Openai, GPT-amge-1 employs the same safety handrails as the generation of images in Chatgpt, including guarantees that restrict the model to generate content that manages the company’s policies. Developers can control moderation allergy, which can be adjusted on “automatic” for standard or “low” filtration for the lowest restriction. Low nomination limits lower categories of inappropriate content, for all Openai documents provided for Techcrunch.

Openai also says that all the images created using GPT-IMAGE-1 are a watermark with C2PA definition data So that it can be identified as created by Amnesty International platforms and applications.

The pricing is $ 5 per million icons for text entry, $ 10 per million pictures of pictures, and $ 40 per million pictures of photos. (Symbols are the raw parts of the data processed by the model.) That translate into about 2 cents, 7 cents, and 19 cents for each image created for low, medium and high -quality square images, respectively, according to Openai.

Openai says that companies, including Adobe, Artable, Wix, Instacart, GodDDY, Canva and Figma, are already used or tried with GPT-amge-1. For example, the Figma Figma design platform allows users to now create and edit images via GPT-IMAGE-1, while InstaCart tests the image form for shopping lists.