OpenAI Reviews 

Dear OpenAI Support,

I am writing to make a formal complaint regarding the reliability and quality of the paid ChatGPT service and to request appropriate compensation or account credit.

I pay a significant amount for access to this product and, unfortunately, I have experienced repeated technical failures and interruptions which have materially affected my ability to use the service for the purpose for which I am paying.

This is not an isolated incident. I have experienced repeated instances of the service breaking down, failing to complete tasks correctly, losing continuity during work, requiring tasks or instructions to be repeated, and generally failing to provide the level of reliability I would reasonably expect from a paid product.

Indeed, while discussing these very problems with ChatGPT and trying to work out how to raise the matter with OpenAI, the service failed again.

My concern is therefore not simply that an occasional technical error has occurred. Software can of course experience faults. The issue is the cumulative effect of repeated failures while I continue to pay the full subscription price.

These failures cost me time. They interrupt work, require me to repeat instructions and recreate work, and substantially reduce the value of the subscription I am paying for.

I would therefore like OpenAI to:

Review the problems and service failures associated with my account and usage.
Review the reliability of the paid service I have received over the relevant subscription period.
Provide an appropriate refund, partial refund or account credit reflecting the periods during which the service has not performed to a reasonable standard.
Explain what OpenAI proposes to do to address the repeated reliability issues I have experienced.

Please do not treat this simply as a request to cancel my subscription. I am raising a complaint about the quality and reliability of a service for which I have already paid and am specifically requesting compensation for the deficient service received.

I would appreciate this being reviewed by a member of the appropriate customer support or billing team rather than receiving only an automated troubleshooting response.

If OpenAI considers that compensation is not appropriate, please provide the reasons for that decision and explain the escalation or complaints procedure available to me.

I would prefer to resolve this directly and reasonably with OpenAI.

I look forward to your response.

Kind regards

The worst customer support experience i have ever encountered in 25 years of working with IT service providers. I don't believe there is an actual human being behind the replies i got. Every reply i sent was met with a complete disregard for my previous reply and at times it could take a week to receive one of these utterly useless replies.

I am extremely disappointed with the way OpenAI manages its usage quotas.

This is now the FOURTH time I have experienced the exact same problem.

My quota operates on a 7-day cycle. On the day the quota resets, I still had around 60% of my usage remaining. Instead of allowing me to use what I had left, OpenAI reset the quota and the remaining usage simply disappeared.

So I am paying for a subscription with a defined usage allowance, yet a significant part of that allowance can effectively be wiped out before I get the chance to use it.

Once could be a technical issue. Twice would already be frustrating. FOUR times is unacceptable.

I am a heavy paying user and I monitor my usage carefully. Losing around 60% of my remaining quota because the system decides to reset the cycle makes absolutely no sense from a customer’s perspective.

If a quota is allocated for a 7-day period, customers should be able to use what they paid for during that period without seeing a large remaining balance simply disappear.

After this happened for the fourth time, I cancelled my subscription.

I don’t like paying for a service and then feeling like I’m being taken for a fool. If I pay for an allowance, I expect to actually be able to use it.

OpenAI seriously needs to rethink how these quota resets work and how unused allowance is handled.

Four times was enough. You’ve lost a paying customer.

Signing up for a purchase and don’t want card details kept for future purchases. Use a regenerative virtual card and purchase blocked by OpenAI as can’t copy card details. So very very untrustworthy. Avoid avoid avoid!!

OpenAI repeatedly suggested that I “try” ChatGPT Work. After years of using Plus, my account had accumulated so much clutter that starting a separate workspace and migrating the chats and projects I actually needed seemed like a sensible solution.
I purchased ChatGPT Work for a year. Almost immediately, I discovered two things I wish had been made unmistakably clear before I paid: I could not migrate my established projects from my personal account into the new workspace as I expected, and Work required me to pay for two seats rather than one.
My total charge was $725.75.
Once I discovered these limitations, I wanted to cancel and receive a full refund. This was essentially immediate—I wasn't someone who had been using the service for months and then changed my mind. I had just purchased it and discovered that it could not serve the purpose for which I bought it.
When I sought a refund, however, I discovered that my purchase was being treated as a business purchase and that I was not eligible for the full refund I expected. I do not recall seeing anything before purchase that made this consequence sufficiently clear to me.
That is what turned a disappointing purchase into a serious loss of trust.
I was not trying to leave OpenAI when this happened. I was trying to expand my use of it. Instead, within days, I went from purchasing a year's worth of ChatGPT Work to thinking seriously about how to move my existing projects elsewhere and eventually leave OpenAI entirely.
This also comes after months of frustration with clutter and the amount of time I have spent reorganizing and reconstructing work inside ChatGPT. I stayed because I have years of projects, hobbies, ideas and other work accumulated here. Now that accumulated work feels less like a reason to stay and more like something I need to disentangle myself from.
If you're an individual considering ChatGPT Work because you want a cleaner environment for work you've already developed in a personal ChatGPT account, verify the migration capabilities, minimum seat requirements and refund terms yourself before paying—especially before purchasing annually.
I wish I had.
I paid $725.75 while trying to deepen my relationship with OpenAI. The experience instead convinced me to start looking for alternatives. Keeping the revenue from one transaction was apparently worth more than retaining my trust and future business.

One star.

In July 2026 I paid $242 for prepaid ChatGPT Credits (Codex). My balance of 491.8 credits was wiped to zero within one second with zero usage in between (recorded in my application logs); after I paid for 1,000 more, OpenAI support confirmed in writing my balance was "still $0". Later purchases burned out in minutes with no spend-rate warning while the service kept asking me to pay again - four times in one evening. When I asked for an itemized record of the consumption, OpenAI refused five separate times in writing: "we're unable to provide itemized credit consumption logs, internal ledger records, or other internal usage data." My GDPR request for my own account records was first answered by pointing me to files on my own laptop, then put under a two-month extension while support kept refusing the same records. I have now filed complaint DPC0826667433 with the Irish Data Protection Commission, with the full correspondence attached. I was scrupulously honest with them - I even publicly corrected my own claims where my logs showed real consumption. They chose five refusals over showing one receipt. If you prepay OpenAI for credits, know that their official position is that you are not entitled to know where they went.

⭐ 1 Star

If you're a small business owner, 🚨 BEWARE 🚨

This is my opinion based on my experience.

I spent months building business systems, manager guides, workflows, documentation, and long-term projects inside ChatGPT. After finally getting everything working the way I wanted, I upgraded from a Personal account to ChatGPT Business because I expected a better experience.

Instead, the migration reported success and then left me unable to access months of my work. My conversations disappeared, eventually reappeared, and now either tell me I don't have permission to access conversations that I created, or they remain stuck on an endless loading screen.

The migration warned that it was permanent. What it did not do was clearly tell me to back up my business-critical work before proceeding. Had I been given that warning, I would have stopped immediately and backed everything up.

What makes this worse is that OpenAI's own help section already documents this exact problem. The information was sitting there the whole time. But at no point during the actual migration did they surface that warning to me directly — instead, I was expected to already know to go dig through their help articles on my own, find the one buried article that described my exact issue, and fix it myself before it happened. A known issue that serious should be a warning shown to you before you click the button, not a scavenger hunt you're expected to complete after your business data is already gone.

Instead, my business workflow is now on hold while I wait for support. Their own email told me it could take days to hear back. Days. That's days I'm not able to touch a very important project I'm in the middle of — not because I did something wrong, but because their migration broke it and their response time doesn't match how urgent this is for a small business trying to actually run on their platform.

This isn't just about one bug. Over the past several months I've dealt with repeated friction: rebuilding context, spending 20–30 minutes every day maintaining ChatGPT's memory instead of doing my own work, product issues, and now being locked out of the very work I upgraded to protect.

As a business owner, I teach my managers that customers leave through death by a thousand paper cuts. One mistake is understandable. Repeated friction destroys confidence.

That's exactly what has happened here.

Based on my experience, I no longer feel that the needs of a small business customer like me are being adequately protected when something goes seriously wrong.

Here's my personal opinion, and I want to be clear that it is opinion, not something I can prove: I believe small businesses like mine simply aren't a priority for OpenAI. My belief is that their real focus is on landing and protecting the big Enterprise contracts — the companies paying for hundreds of seats — and that those customers get taken care of. I believe the rest of us, the small business owners paying for a handful of seats, are treated as an acceptable loss. Churn we represent is, in my opinion, just a built-in part of their business model rather than something they're actively working to prevent. I can't prove that's their strategy, and I want to be fair and call it what it is — my opinion, formed from watching how a serious support failure like this one gets handled for someone my size. But it's a strong enough opinion that it's changed how I plan to evaluate vendors going forward.

Whether you're considering using ChatGPT for business or already rely on it, I strongly recommend keeping independent backups of everything you can't afford to lose and having a contingency plan if your access is interrupted.

The AI itself is impressive. The platform experience has convinced me to seriously evaluate moving my business to another AI provider.

I created PR to my codebase. Asked for review.

Refused because of security guards like if I am hacker.

Asked to incorporate or get certification to do code review.

This is unacceptable for force me to pay other people to make my app secure. Is not ai was created to avoid that?

I am extremely disappointed with OpenAI’s customer support and billing practices.

I accidentally ended up paying for two ChatGPT Plus subscriptions—one through Apple and one directly through OpenAI. I only intended to have one subscription.

I contacted support because I wanted to keep my Apple subscription and have the OpenAI subscription canceled and refunded. Apple could not help because the duplicate charge was billed directly by OpenAI.

The biggest issue is that I never used the OpenAI subscription. I didn’t even know what email address it was linked to. I had to ask support to identify the account because I had no recollection of which email it was associated with. I couldn’t log into an account when I didn’t know what account it was.

Despite explaining that this was an accidental duplicate subscription, OpenAI refunded only $4.91 of the $21.74 charge, stating that I contacted them 24 days after the purchase. That completely misses the point. This wasn’t a case of deciding to cancel after 24 days—it was a duplicate subscription that I never intended to purchase in the first place.

I expected a company like OpenAI to make this right by refunding the duplicate charge. Instead, I was left paying for a subscription I never intended to have and could not even access because I didn’t know which account it belonged to.

This experience has seriously damaged my trust in OpenAI’s billing and customer service.

Open A.I. has resorted to scamming people with double charging and then getting the A.I. "help" to gaslight you into thinking you use devices you do not even own and claimed a second subscription.

you can't do anything like you type 2 messiges with chatgpt and thay put you on a limit becase thay are greedy and you it cant generate basic request and than thay hit you with a 8 hour limit and pushes you to buy there premium becase thay are greedy

works ok but donot use codex : Paid subscription, poorly explained monthly cap, unpredictable consumption, €100 surcharge offered once the project is started, and no proper visibility before being blocked.so it's a scam

I contacted OpenAI Support about a separate credits purchase that had not appeared in my account. Support said it had refunded that purchase, but then also cancelled my paid ChatGPT Pro subscription without my request or permission.

I immediately explained the mistake and repeatedly supplied receipts, account information and existing case references. Instead of resolving it, I received generic AI-assisted replies, contradictory statements about my payments and several new case numbers that appeared to ignore the existing complaint.

At the time of writing, my paid service has not been restored, the matter remains unresolved, and I am having to pursue recovery through my card provider.

The technology may be advanced, but the billing administration and customer service have been chaotic, impersonal and completely unacceptable.

Would give no stars if possible.
OpenAi has been taking money of my bank account even though I dont have or use any of their products!
It is very scammy!!
After multiple emails back and forth and providing the same info over and over again they are still completely incompetent to help.
Had to cancel my card to ensure them not deducting more.
Avoid at any cost, absolutely no customer service whatsoever.

Keeps improving at least in the same pace like other AI models. Usage limits are better compared to alternatives, so I don’t regret subscription at all.

The AI assistants are absolutely useless and they apparently don’t have the option to talk to a real person on the phone. Absurd.

I have no idea that it would get so many things wrong. I've wasted 8000 units just in the last two hours and it got every single thing I asked wrong. I have two people talking, for example, and they're talking at the same time saying the same words and it won't stop doing that no matter how many times I render a new verge, horrible

it is not ai, that claim is false, it is just a very extremely sophisticated algorithm that can talk but sucks at any code writing, results are poor, you never get what you want if it is object programming, this chat bot can only talk and talk is cheap

I signed up expecting to try the service, but I was unexpectedly charged €29.99 for a subscription.

The checkout process was not transparent enough about the recurring payment. I do not believe the subscription terms and pricing were presented clearly before the payment was authorized.

After using the service, I found the usage limits to be far more restrictive than expected, making the service significantly less useful than advertised.

After reading other reviews on Trustpilot, I realized that many customers have reported similar experiences with unexpected subscription charges and unclear billing practices. Based on my experience, I cannot recommend this service.

I accidentally purchased a ChatGPT Plus subscription for $20 on July 12, 2026. Less than one hour later, I canceled the subscription and requested a refund.

I live in Italy and repeatedly explained that I had contacted OpenAI within the 14-day period applicable to consumers in the European Economic Area. I requested either a full refund or at least a prorated refund.

OpenAI Support reviewed Cases 11354938 and 11860533 but repeatedly refused the refund. They did not provide any specific contractual clause or legal basis for the refusal. I also asked for the details of the responsible legal entity and the official contact for formal consumer complaints, but support claimed that it did not have verified information available to provide those details.

Instead of answering my questions, support repeatedly sent almost identical template responses. They acknowledged that I live in Italy and that I requested the refund less than one hour after purchase, but still refused both a full and prorated refund without explaining why the published refund provisions for EU consumers did not apply.

The subscription was canceled and will not renew, but the $20 was not refunded. I am now preparing a complaint to ECC-Net Italia.

Very disappointing handling of a simple refund request and extremely poor transparency from a company of this size.