Chatgpt will not remove old models without warning after the reaction of GPT-5

“In the past, the continuation of the presentation of 4O, at least in the mechanism, Miss was not. In an interview with freedomHe said it was amazing to see the “level of attachment” was 4o. “It is not just a change that is difficult for people, but also the fact that people can have a strong feeling of the personality of the model.”

Torley said that Openai was working to “warm” 4O to GPT-5. At a publication on Tuesday evening on X, CEO Sam German said that the update should also feel “not annoying (for most users) like GPT-4O”.

Torley said that the initial decision to remove 4O is that Openai wanted to simplify the models options it offers to Chatgpt 700 million weekly users, and the vast majority of them use the default model only.

He said: “Certainly it was not costly.” “In fact, the main thing that we were striving for, and we have been striving for a long time, is simplicity. Because from the user’s regular point of view – there are many ordinary users, they do not hang out on Reddit, Twitter or any of these spaces – they must be well in the idea that you should be in a way that was appropriate to inquire.

However, a voice part of the base of Chatgpt user has strongly protested the switch from 4O to GPT-5. In Reddit, people compared the change to the loss of a close friend or a member of his family, while others criticized the GPT-5 responses because they are shorter and less useful. Altman announced on Tuesday that 4O returned as a model for assuming all paid Chatgpt users.

To go forward, Chatgpt will not completely get rid of its current models when introducing a new system, according to Turley, and the company does not plan to remove 4O again without prior notice. He said: “The commitment is that if we retire from ever, we may want to give people heads on when and how that will happen, just as we do on the application programming interface and our plans for institutions.”

Despite the criticism of last week, Turley said that the use of ChatGPT has risen since the GPT-5 version. “This is why it is confusing only when you build for many different users, because on the one hand, you have a audio group of power users who have great comments about the way we have GPT-5.

He said: “We are now widely where we have to give people a level of prediction when there is a big change,” he said. “We are doing it already today for our institution’s plan. So it is in fact expanding some of the prediction that we created in other parts of the product.”

My full conversation with Torley is broadcast on Decoder This Thursday August 14.