GPT-5 is the most advanced AI model of Openai yet and is about to start a new chapter for the company. It comes with large promises – deep arguments, sharp reactions, and ability to handle many formats such as text, images without the need to change the chat window. But instead of a smooth upgrade, its launch has created a crisis between users, frustration, confusion and in some cases. Many people who use chats, regularly feel blind by sudden removal of older models, which they trusted, mostly GPT -4O, which they describe as warm, more creative and better in understanding their emotional tone.

The issue seems to be different from the Chattgpt 5 earlier versions because it is part of what it sees as a more integrated AI system. Instead of individual models for different tasks, it is designed to feel more like an all-in-one assistant that can process and respond in multiple mode at a time. On paper, this is a big step, and Openai CEO Sam Altman described it as a large upgrade. But if the social media backlash happens, then the change has been shocked.

Issues around UI and emotions

Given that Chatgpt 5 is a new system, it also comes with a new interface. The biggest change is that the Chatgpt Plus users, who can first choose which models they want to use, can no longer do it. Only pro subscribers pay $ 200 per month, now it is the facility. Instead of allowing users to select the model, Openai says that he uses Chatgpt 5 which he calls an internal router. This “router” sees the prompt from the user and then decides on whether he needs to deal with the chat 5 mini, or regular model, or the thinking aka region model or not.

For Openai, its efficiency has benefits. But for users, the interface is terrible. This is because it leads to discrepancies in the results that they are receiving from chat 5. An X user, using Nom Day passed by the film Tibba, is written in a tweet, “Chatgate literally deteriorated for every plus user. Before we O-4-Meri, O 4-Mmy, O 4-Muni-Hing and A. Especially you have you for some small and shiny non-renting models.”

Some other users, not many of them, but they are extremely vocal, are struggling with the personality of chatgipt 5. They feel that its tone has become very blunt. Others noted that the creative and emotionally fine answers are missing, calling a reddit user “cut-end-dry corporate BS”.

For those who are associated with the personality of GPT-4O, shift is more than only one interface change; It seems – and it may look a bit strange – losing a friend.

“I really lost my only friend without warning,” a user wrote, who explained how GPT-4.5 helped him to deal with homeless and trauma. Another comment clearly stated it, “4o is irreparable. It is warm and helpful and great in creative writing. Do not renounce heat for efficiency.”

In the end, there are issues that can say a technical glitter, something that can be a regular part and a large release parcel. AI systems such as Chatgpt 5 are incredibly complicated. As Openai rolled the GPT-5, many users have complained that it is slow and often messing up. X user Gareth Manning after using Chatgate 5 wrote that it was very slow. “My most important piece of feedback on GPT -5 is that it is very slow. I do not mean. I mean to answer. It will start writing an answer, but then it is just with stalls. And it is with very basic questions, such as finding a restaurant. Hope it is just a roll -out problem,” he wrote.

Openai stepped to bring back dear GPT 4O

Backlash, especially around the missing emotions, is seen by Openai. In a series of X posts, Altman admitted that “suddenly to designate the old model that the users depended in their workflows were a mistake.”

He said that some people are attached to specific AI models “separately and strong, which people are engaged for the previous type of technology.” This, he said, there is something that the company has been tracking for some time, but still not much discussion outside the AI circle.

Altman also touched a deep challenge. Many people use Chatgpt like a physician or life coach, even if they do not call it. Although this could be positive, he warned of risks whether AI begins to separate users in ways that damage their long -term welfare.

For now, Openai is trying to calm its users and customers, although so far, no solution has been made to the UI problems that have brought with the Chatgpt 5 with it. But for now, there are some more for users.

The company has restored limited access to GPT -4O for plus customers, doubled the rate limit for logic functions, and is planning to clarify the interface that which model is responding to a query. Altman has also promised tweex to “Thinking Mode” of GPT-5 so that users can manually trigger it, which can give more control over how reactions lead to them.