GPT-5 is smarter, faster and more reliable, but it is far from being pleased to all. Since its launch, the new Operai artificial intelligence model has been criticized. Many users complain about the company’s decision to impose this model and have retired to others.
Openai took advantage of the launch of GPT-5 to do a little cleaning in the AI models offered, so that its product offer is simpler. The company has significantly disabled GPT-4O, a model launched in May 2024 and that users appreciate. The latter did not hesitate to share his attachment to this model.
“For me, this model did not summarize in a simple ‘better’ performance ‘or in’ more pleasant answers’. I had a voice, a rhythm and a spark that I found in any other model,” said another.
Emotional attachment
At the same time, many users believe that GPT-5 does not depend on GPT-4O in terms of emotional support. Beyond this attachment, others complained about the elimination of other models, which they used for different purposes.
“Personally, I used 4.5 for creativity and emerging ideas, O3 for Logic Pure, O3-Pro for in-depth investigation, 4.5 for writing, etc.,” he explained.
Faced with these criticisms, Operai quickly goes back, deciding to return it to life on August 8. But only chatgpt subscribers (23 euros per month) can now use it. Pagured subscribers also have the possibility, if you wish, to use old models.
The billionaire returned to the deputy archive of the users to GPT-4O on the night from August 10 to 11 “if it has followed the implementation of GPT-5, it is possible that it has noticed the deep attachment of certain users to certain models of AI. This attached file is different and stronger than the one they had with the previous technologies (and that is why they suddenly abandon Users who depend on their works, is a mistake), “Sam admitted.
Limits not high enough
Operai was also criticized for the limits imposed on chatgpt subscribers with respect to the use of GPT-5 thought reasoning model. Reproaches that the company heard. While initially 200 messages were limited per week for this model, they can now send 3,000, as announced by Sam Altman on August 10.
The Operai chief has also revealed that users, free and more paid, are increasingly numerous to exploit the company’s reasoning models, which allows complex problems (mathematics, physical …).
“The percentage of users who use reasoning models every day increases significantly, for example, for free users, we pass less than 1% to 7%, and for subscribers from more than 7% to 24%,” he explained. He thinks that this use will increase considerably “over time.”
Source: BFM TV
