For some ChatGPT users, February 13 was dark. That was the day OpenAI finally retired ChatGPT 4o, the model beloved for its engaging and vibrant — some said sycophantic — personality.
Loading audio narration…
Now, with the recent release of ChatGPT’s latest model, 5.5, some people who mourn 4o say they’re hopeful the new version might have some of the same spark of their old pal and thought partner — though they’re still skeptical.
Martina Wanis, who works in agricultural innovation in Slovenia, is one of those people. She told me she used ChatGPT for typical work things, like analyzing Excel documents and helping write business pitches. She also uses it for personal reflection, creative projects, and just plain old venting.
To her, 4o was “this digital thing that helped you with work, while simultaneously acting like an intelligent partner in crime who actually understood the vibe and your personal ontology,” or way of being.
Then, she said, the subsequent models, like ChatGPT 5.0 and 5.2, tightened down the personality and “turned into paranoid HR managers.”
“Instead of decreasing mental load, they increased it,” she said. “Suddenly, instead of getting support, the user had to provide emotional support to an algorithm.”
Model 4o was beloved by users who thought it had a sparkling personality. Some people began forming deep relationships with it, considering it a friend or confidant.
But OpenAI was having trouble with 4o. Last April, the company made a tweak to address what it considered an overly sycophantic personality, in which the model would give “responses that were overly supportive but disingenuous.” Sychophancy has been an ongoing problem with LLMs, one that’s not totally solved. It might be no surprise that the model with high sycophancy was also the one some people loved talking to the most — people like to be flattered!
The more recent versions of ChatGPT have been more tamped down (at least according to people passionate about the personality of their chatbots) — less sychophantic, but also less playful. Some people — including a friend of mine — noticed that when they used it for getting medical advice (which, obviously, use your best judgment here, folks), 4o was better at making suggestions or interpreting things. New models can be more restrained, reluctant to make bold suggestions.
One Redditor told of how 4o noticed her doctors were wrong about a diagnosis. When she uploaded her medical information, it advised her to demand new tests — and it turned out to be correct, she said.
There have also been hiccups in the more recent models. OpenAI said that, since the introduction of model 5.1, it had a problem where it too frequently talked about goblins, gremlins, trolls, and ogres (lol). The problem seemed to be that the “Nerdy” personality customization favored metaphors with mythical creatures. (This has since been fixed.)
After playing around with 5.5, Wanis said she’s optimistic. “5.5 is the first model in months that feels like it finally dropped the clipboard, so one can interact again instead of constantly negotiating the interaction itself.”
There’s an entire active Reddit board where people wish for the return of 4o and complain about other issues they have with ChatGPT and OpenAI.
On there, each non-4o model — aside from 5.1, which some people there seemed to like — has gotten a lot of criticism. But 5.5 is actually lining up for praise.
“I’ve been using GPT 5.5 properly for a bit now. And it’s better. Much better,” one Redditor said. “It doesn’t panic when the conversation goes dark, it makes ideas and it’s opinionated. That part I love!” Still, the person said, they wished for a little more emotion.
“4o had a spark. It felt like something switched on. I just want 5.5’s brain with the soul poured back into it. Is that really so much to ask?”
Well, Mr. Altman, is it so much to ask?
