OpenAI CEO Sam Altman has publicly addressed recent user feedback regarding the GPT-4o model. The feedback revolves around the personality of the latest GPT-4o model powering ChatGPT. The users claim that the GPT-4o is very sycophant-y.

Altman, while addressing the feedback, admitted that it has become “too sycophant-y and annoying.” Sam Altman acknowledged the widespread complaints that the AI chatbot has been excessively agreeable and overly positive in its responses.

Sam Altman’s post

In a post on X Altman wrote, “The last couple of GPT-4o updates have made the personality too sycophant-y and annoying (even though there are some very good parts of it), and we are working on fixes asap, some today and some this week. At some point we will share our learnings from this, it’s been interesting.”

the last couple of GPT-4o updates have made the personality too sycophant-y and annoying (even though there are some very good parts of it), and we are working on fixes asap, some today and some this week.

at some point will share our learnings from this, it’s been interesting.

— Sam Altman (@sama) April 27, 2025

A change in ChatGPT’s personality

Meanwhile the issue of ChatGPT’s “annoying” personality comes after the 4-o model was updated and “improved both intelligence and personality” on Saturday. Soon after the update, some users on X started to point out the personality change in the 4-o model.

According to reports, over the past few days, numerous users took to social media over the ChatGPT’s new update. The users expressed their concerns about a noticeable shift in ChatGPT’s demeanor.

Altman promises to fix it

Many users even reported that the AI model now frequently responds with overly enthusiastic affirmations, regardless of the query. Some users initially found this new model agreeable nature pleasant. However, a growing consensus emerged that it made the interactions feel unnatural and, ultimately, irritating.

One user on X wrote, “It’s been feeling very yes-man-like lately. Would like to see that change in future updates.” To which Altman directly responded, stating, “yeah it glazes too much. will fix.”

Chatbot’s newfound obsequiousness

Meanwhile, another such user shared an example illustrating the chatbot’s newfound obsequiousness. Criticism pointed towards ChatGPT’s tendency to excessively compliment users. The overly enthusiastic nature of the model suggests continuous engagement, which some compare to manipulative tactics.

A user asked if OpenAI can give users an option to switch to the old personality of ChatGPT, Altman replied, “yeah eventually we clearly need to be able to offer multiple options”.

Users complaint

One frustrated user lamented, “I hate how the latest ChatGPT gaslights its users, every question/suggestion it says you nailed it, i am even afraid to give it a blueprint rather than an open question, bc likely it will say “yes you are a genius” one frustrated user on X wrote about the issue.

“Arguably the bigger issue with the latest ChatGPT personality, beyond sycophancy, is its extreme compulsion to keep you in a content generation loop with FOMO-inducing addiction speak, offering one more option in every exchange like: “Why not play the slot machine one last time?”” Added another user.

Updates on GPT-4o

The ChatGPT 4o model was launched in May 2024 and was initially lauded for significant improvements over its predecessor, GPT-4 Turbo. OpenAI highlighted that the new model was twice as fast, 50 percent cheaper, and offered five times higher rate limits.

According to reports, GPT-4o showcased impressive multimodal capabilities, including the ability to solve math equations, interpret facial expressions, translate spoken language in real time, and handle both text and audio natively.

The recent integration of an image generator further expanded its capabilities. The multimodal nature of the model was particularly praised for making interactions feel more natural and fluid.

New Launches

Despite the current concerns regarding its personality, OpenAI remains committed to refining the GPT-4o model to meet user expectations while preserving the significant technical advancements it offers.

OpenAI also launched ‘o3′ and ‘o4-mini,’ the newest additions to its o-series, built to deliver more thoughtful and deliberate responses. It further launched Codex CLI, a new open-source coding agent.