<img src="https://media.daily.dev/image/upload/s--qpmoBKyp--/f_auto/v1757520458/ugc/content_45da240c-c1ac-4dc5-a7e0-80df2e606fe7?_a=BAMClqZW0" alt="image">
Unless you’ve been living under a rock, ChatGPT-5 sucks, hardcore. Current criticisms against ChatGPT-5 largely center on its perceived lack of improvement and a shift in user experience, which many feel is a downgrade from its predecessors. The backlash from the user community, particularly paid subscribers, was so significant that OpenAI had to restore access to older models like GPT-4o.
One of the main complaints is that the new model’s responses are shorter, blander, and lack the personality and “charm” of previous iterations. Users report that the model feels less engaging and its writing is more formulaic. Furthermore, despite being touted as a “PhD-level” intelligence, critics argue that ChatGPT-5 still makes simple errors and struggles with basic reasoning, especially in areas like math and symbolic logic. The model is also said to still suffer from hallucinations, where it confidently provides false or made-up information. The removal of the ability for users to choose between different models and the implementation of an “autoswitcher” that defaults to a lighter, less capable model for simpler tasks has also been a point of contention, leading many to believe that the upgrade was a cost-saving measure for OpenAI rather than a genuine leap in capability.
Beyond the immediate user experience, the rollout of ChatGPT-5 highlighted broader criticisms about the direction of large language models (LLMs) and OpenAI as a company. The forced removal of older, preferred models and the introduction of stricter message limits for both free and paid users sparked outrage among the community, who felt that their access and control were being stripped away. This move, combined with the lackluster performance, led many to question whether LLMs have hit a plateau and if major breakthroughs are becoming less frequent.
Nevertheless, the obvious reduced quality in product that OpenAI rolled out to the public indicates that they are most likely reserving their best for their high-roller clients such as the United States government or other large enterprises.

Josh M.

ChatGPT-5 has received significant criticism from users who report that responses are shorter, blander, and less engaging than previous versions. Despite being marketed as PhD-level intelligence, the model still makes basic errors in math and reasoning while suffering from hallucinations. OpenAI's removal of model selection options and implementation of an autoswitcher has frustrated users, leading many to believe this was a cost-saving measure rather than genuine improvement. The backlash was severe enough that OpenAI restored access to older models like GPT-4o.

GPT-5 is Trash.

Regardless of gpt5 sucks or not, the big mistake was to remove immediately all other models. This led to lots of frustration and backed fire. It’s an AI market currently so they can do what they want with no consequences

<blockquote>
One of the main complaints is that the new model’s responses are shorter, blander, and lack the personality and “charm” of previous iterations.
</blockquote>
the sycophantic tendencies have gotten a lot better though. I’m wondering if that is the driving factor here. That is, people wanting the sycophantic answers.

I actually go in Duck Duck Go’s website, then open <a href="http://Duck.ai" target="_blank" rel="noopener nofollow">Duck.ai</a> and ask what I want, because it uses GPT 4o-mini and that’s better than GPT-5. It is sad that I have to use “tricks” to use GPT-4.

AI will be AI, it’s the best in being a sucker.