The launch of OpenAI’s GPT-5 on the 7th of August was meant to bring improvements in writing, coding, healthcare advice and visual tasks. Instead, the update has left a portion of the user base unhappy, as it replaced all earlier ChatGPT models in one go.
Among the models removed were GPT-4o, GPT-4.1, GPT-4.5, GPT-4.1-mini, o4-mini, o4-mini-high, o3 and o3-pro. OpenAI said GPT-5 was its most advanced system so far, able to give quick answers when needed and longer responses when more depth was required. It also came with four new optional personalities which are Cynic, Robot, Listener and Nerd, as well as an improved reasoning engine called GPT-5 Thinking.
While the upgrade was promoted as a step forward, many users felt the changes hurt their experience. On Reddit, one person said losing GPT-4o was “like killing a friend of millions of users”. Another wrote that it had helped them through depression and anxiety, describing the model as warm and understanding in a way GPT-5 was not.
What Has OpenAI Decided To Change?
OpenAI’s chief executive Sam Altman responded within days. Writing on X, he said the company had underestimated how important certain qualities of GPT-4o were to people, even if GPT-5 was better in most areas. He confirmed that GPT-4o would return for ChatGPT Plus subscribers, giving them the choice to keep using it.
Altman also told users in a Reddit AMA that OpenAI would track usage to decide how long GPT-4o stays available. He added that GPT-5 would start performing better immediately after a technical fault with its autoswitcher had been fixed.
This decision is a change from OpenAI’s earlier plan to replace every old model. Altman has said that personalisation is a priority for the company, with the long-term goal of letting people adjust the style and tone of the AI to suit their preferences.
More from News
- Two Thirds of Brits Are Telling AI Chatbots Their Secrets But Don’t Trust The Companies Behind Them
- Video Game Companies Are Abandoning “Buy Our Console To Play This” – Here’s Why
- UK Gambling Operators Face New Deposit Limit Rules From 30 September
- How A Ransomware Gang Hijacked A Rival Gang’s Website – And Is Now Extorting Them For “2.333% Of Their Net Worth
- Meta Promised Unmatched Privacy For Muse – So How Did Researchers Crack It So Quickly?
- Digital Avatars Are Bringing Deceased Artists Back To The Stage – Where Do We Draw The Line?
- Can Europe Solve Its Critical Minerals Problem Fast Enough To Support Its Tech Ambitions?
- Google Has Agreed To A £260 Million Settlement: Who’s Eligible For Compensation?
Why Do People Prefer GPT-4o?
Although GPT-5 has stronger abilities in coding, mathematics and health queries, some users find GPT-4o more consistent in conversations. It is described as having a tone that feels more relatable, while GPT-5 can seem less personal.
Altman has said different people want different traits from ChatGPT. Some prefer an analytical style, while others want more warmth or emotional intelligence. GPT-4o appears to have struck that balance for many, making its removal feel abrupt.
For certain subscribers, GPT-4o was needed for their working. One Reddit post said everything that made ChatGPT useful for their work had been taken away without warning. In response, that user cancelled their paid subscription.
How Is GPT-5 Being Released?
According to Altman, GPT-5 has already reached all Pro users and is close to reaching every account. The company is working on changes to make the model feel warmer in conversation, although more advanced customisation features will take longer to develop.
The update introduced voice mode for custom GPTs, raised usage limits for both free and paid accounts, and allowed speaking instructions to change depending on the situation. GPT-5 merges the abilities of older versions into one model, meaning users no longer need to pick from a list of numbers.
