March 28th.OpenAI CEO Sam Altman said on his X-account on March 27, local time.ChatGPT ofVincent Figureapplication demand is too high, resulting in GPU overload.The company will work to improve efficiency and introduce some frequency limits to the ChatGPT image generation feature for the time being.
Altman humorously said, "It's really fun to see how much people love the image features in ChatGPT, but our GPUs (graphics processing units) are 'melting'."
Altman did not explicitly disclose the specific frequency limitation criteria, but said he "hoped" the restriction would not last long.
Oltman also mentioned thatFree users will "soon" be able to generate up to three images per day.
1AI notes that Altman also said, "In addition, we have rejected some generated content that should have been allowed, and we are working on fixing these issues as soon as possible."
OpenAI has upgraded and optimized its image generation tools with its GPT-4o model. The new model produces images that are often more lifelike than ever before, and also makes significant strides in past technical bottlenecks such as text rendering. In an interview with The Verge, the company described the new model as a "quantum leap" over previous models.
Earlier today, OpenAI announced an update to the GPT-4o model in ChatGPT, with updates to optimize multi-instruction parsing, improve parsing of complex technical and programming problems, improve creativity, and more. The updated GPT-4o is available to all paid users, while free users will have to wait a few weeks.
