According to CEO, Altman himself, the enthusiasm around the more available (and more advanced) possibilities of generating the image of chatgPT forced OpenAI to “temporary” speed limit limiting. “This is a great cool vision of people who love paintings in chatgpt, but our GPU is melting” Published today on X. Altman did not determine what the rate limit is, but he said that the security “I hope” would not have to be in place, because Opeli is trying to augment its efficiency in the avalanche of demands.
Krut demand has already caused that the artificial intelligence company has rejected the availability of a built -in image generator for users at a free chatgpt level. But apparently this remedy itself was not enough to alleviate the stress associated with the OpenAI infrastructure. (Altman said that free users will be able to generate for three photos a day).
Opeli uses its GPT-4O model for the newly improved image generation tool. The resulting works are often even more realistic than before – although you can take it in any number of directions – and the up-to-date model has also made progress with past obstacles, such as rendering text. In an interview with The VergeThe company described this as a “change of steps” above earlier models. But limiting speed is another reminder of how much technological power (and energy) is required to generate images in chatgpt.