Several users have reported issues with ChatGPT worldwide on Wednesday, including in Israel, with nearly 1,500 reports submitted on the specialized website DownDetector in the last 24 hours.

Users are reporting a noticeable delay in receiving answers and, in some cases, a complete lack of response from the chat. Israeli users are also experiencing similar difficulties using the tool.

As of Wednesday morning, OpenAI has not yet released an official response regarding the cause of the glitch; however, the problem has affected many users worldwide.

Last April, a similar problem was reported in ChatGPT, where the system presented users with the error message "Something went wrong" despite repeated attempts to refresh the page.

In the past, OpenAI CEO Sam Altman has addressed the challenges of high demand and system load, hinting that the company may not have enough computing capacity to handle the heavy load generated by the increasing use of the tool. This statement also refers to the AI-based image creation tool, which he claims is "melting" the company's computers due to high demand.

Downdetector notified problems with OpenAI's ChatGPT.
Downdetector notified problems with OpenAI's ChatGPT. (credit: screenshot)

New ChatGPT-5: What does it have?

The situation happened just weeks after the latest version of ChatGPT was launched by OpenAI. Currently, a free, limited version is available to everyone, while the paid subscription offers unlimited access.

According to the company, this version will be smarter, faster, and more widely useful than its predecessors, with dramatic upgrades in writing, computer programming, and medical applications.

Altman explained during a briefing that when he tried to return to GPT-4, the experience was "shocking" due to the significant upgrade felt at every level. “It is like working with a team of doctors from different fields at any given moment," he said.

The main upgrade in this new version is focused on reducing the rate of “hallucinations” and improving the information quality that the AI model can provide. In other words, the generation of false information, a known issue with artificial intelligence, should be drastically reduced.