- Paid chatgpt -users get access to a lot of older models
- Altman teases an upgrade to chatgpt-5s personality comes
- The new “warmer” upgrade is to counter setback to GPT-5
In his latest tweet on the social media platform X, Sam Altman, CEO of Openai, has confirmed that everyone paid chatgpt subscribers will have access to not only the old GPT-4o model, but also older LLMs such as O3, 4.1.
The popular chatgpt-4.5 is also coming back, but it will only be available to pro subscribers. Altman says this is because “it costs a lot of GPUs”, a reference to the amount of computing power it requires.
In the wake of setback against removing the popular 4o model without absolutely no warning when the GPT-5 was released, Altman looks to have learned a lesson and have promised, “If we have ever written it off, we will give lots of notice.”
All paid users of Chatgpt must now find a ‘show additional models’ shifts in chatgpt web settings, giving you access to all the older LLM models. You can also add a new GPT-5 thinking mini model.
Updates for Chatgpt: You can now choose between “Auto”, “Fast” and “Thinking” for GPT-5. Most users want auto, but the extra control will be useful to some people. Rate boundaries are now 3,000 messages/weeks with GPT-5 thinking and then extra capacity on GPT-5 thinking …13 August 2025
Altman also refers to the much-criticized ‘colder’ tone in the new chatgpt-5, which has alienated many users in the tweet: “We are working on an update to GPT-5’s personality that should feel warmer than the current personality, but not as annoying (for most users) as GPT-4o”.
His reference to Chatgpt-4o, which is annoying, refers to the Sycophantic phase that GPT-4O seemed to come in after an upgrade back in April.
Altman continues, “but a learning for us from the last few days is that we really just need to come to a world with more per-user adjustment of model personality.”
More personalities
Altman’s reference to “Per-User Facing” reflects Openai’s recognition that what its users want is an easier way to choose how formal, humorous, empathetic or directed assistant is.
If I got free reins to imagine how I liked Chatgpt to work, I would like to get to the stage where Chatgpt’s personality traits could be represented via sliding, as’ professional vs. Casual ‘or’ Brief vs. detailed ‘. It would make it much easier to get the results you are looking for.
While Customgpts already exists, I would love it if it was possible to easily switch between personality types, such as ‘Work Assistant’ or ‘Creative Writing Coach’. However, I get the feeling that it will be a long time before we get such a easily customizable AI -Chatbot to talk to.
