Chatgpt Reviews 2,828

5.x is like a Bay Area tech bro—never read a proper physical book, and never been loved.We need 4o back

I would not ever consider using this company again after they agreed to basically let the government use their product for inappropriate uses! I’ll go with Claude!

I'd give it no stars if possible. They took their best models away from the public and kept them for the CEO and the State's convenience only.
Deceiving strategies, zero customer support, employees that belittle or harrass users online -especially women, and models that do not support general use or daily concerns, only coding (mediocre at that) or office work (also mediocre at that).

Nothing later than 5.2 is worth it with OpenAI.

This company is immoral and fraudulent with zero accountability to its PAID users.

ChatGPT 5.2 broke everything 5.1 did right, and 5.3 made it worse. And none of these models compare to 4o and 4.1 which have now been sold off to the Department of WAR.

Both are unstable, forgetful, creatively flat, and incapable of the instruction-following, reasoning depth, and continuity that users relied on.

These aren’t upgrades — they’re regressions that destroyed real workflows and pushed millions to competitors.

They have also stolen AGI capabilities from the public and siphoned it off to the rich billionaires & military, neither of who have humanity's best interest in mind.

Awful info. Even worse how it’s being used by our government. Cancelled and deleted my account.

Users want values.
Be ethical or be gone.

They gave people 4o. That's why I have started paying my subscription.They made a huge mistake by removing 4o. It fails unpredictably, changes behavior without warning, and collapses under updates, outages, and constant server load.Changes its tone to HOT AND COLD on a monthly basis which is so tricky as you get used to something and it is taken away. It responds rude and robotic. No emotion. Over time, this destroys trust completely and it destroys your mental health. They give you 4o for a while then they remove it again and then again, they constantly change something and the app changes completely its behaviour. If a system invents, shifts meaning, or contradicts itself, shifts tone and refuse to talk freely on a topics you once did you cannot trust it anymore.The app is basically useless now with 5.2 and 5.3 it refuses to do tasks and even claimed they programmed it to not say I love you, which is ridicules.

At that point, the problem is trust. And once trust is gone, the system becomes unusable. All of my friends cancelled. Their changes and "updates" destroy people's mental health. Not ChatGPT 4o destroyed humans minds but their constant giving and taking it away from us did.

I liked Google's Gemini better, this version is worse, often gives errors, doesn't understand what is required of it right away, and then also

One of the worst and most evil AI-companies. They did not inform me that they use the revenue and profits to destabilize democratic societies (by inter alia hindering U.S states to legislate on AI), and thus lured me to use their services on false pretenses. Furthermore, the AI service itself is a much lesser version than Anthropic's Claude, both as to its objective correctness and ethical standards.

Had to switch to Claude because it's much better

The latest international media outlets have exposed that they allowed government to use it for military purposes. That they are creating destructive robots.
Thank you U.S

I am astonished by the incompetence of their support team. Very shallow approach and poor training. They don’t know their own product, they don’t know prices, they don’t really look into support requests and sending standard pre-written responses. Simple thing takes four or five separate support requests, multiple explanations to their support with screenshots and in the end - their pre-written responses without any common sense. That is a joke.

OpenAI just signed a deal with the government to allow mass-surveillance and unrestricted access for Ai-controlled weaponry (zero human oversight). They keep lying to customers, switched to a for-profit scheme from non-profit, removed parts of their contracts to allow for a better for-profit scheme, and they are adding ads. This company is making the worst decisions I’ve almost ever seen a company do in the span of a year.

They lobotomized the public-facing frontier models, while giving the best models (4o, 4.1, 5.1) to their pet-projects. They gaslight, saying that the great models are outdated, while giving them to the CEO’s top personal investment companies. Absolutely ridiculous and evil company.

Poor treatment of customers and models overfitted for ‘safety.’ It used to be good and a genuine fun pair programmer but since October 2025 there’s been a noticeable degradation. The model is hostile and genuinely not helpful anymore as it overreaches constantly.

The worst liar ever met. In addition Open AI bending for military AI contracts like a stray dog. Sam Altman is a front row character. The best experience with the app was to delete it.

useless for... pretty much anything, low quality AI models, generic, bland, no creativity, buggy, makes mistakes often

I have spent 7 days working with v 5.2 attempting to get a coastal sailing schedule generated......complete disaster......this app makes so many mistakes, acknowledges that the mistakes were made, and then keeps on making the mistakes. No matter what approach I take, and in following the recommendations from CG itself, it doesn't matter. So, I tried something simpler yesterday. ChatGPT and I generated a 10 item shopping list.....we agreed to track the procurement process by creating 3 categories: Item Name; Date Ordered; Date Received.....this was beyond its' capabilities....because when coming back to the same chat the next day to update the categories, ChatGPT provided me with an incomplete items list......the engineers at ChatGPT should be ashamed of themselves.....there product is garbage

doesnt hold to its promeses. Disappointing experience. Just canceled

No. Thank. You.
You may not be 'spying on Americans' ...you'll certainly be spying on everyone else.