The $200/month Pro Disappeared, Even the CEO's Account Was Lost

Have you ever imagined paying $200 a month for a ChatGPT Pro subscription, betting your entire workflow on it, and then one morning you wake up to find it telling you—you're a Free user?

This isn't a hypothetical.

Today, OpenAI pulled off what might be the most absurd Bug in AI history: globally, ChatGPT subscription tiers went massively haywire. Pro users woke up to find themselves downgraded to Free, Plus users suddenly gained Pro permissions, and Free users inexplicably got upgraded to Plus.

Even OpenAI's CEO Sam Altman wasn't spared—his Pro account was lost too.

One Flight, All Boarding Passes Printed Wrong

Imagine this scene: You board an international flight, settle into first class, and just before takeoff the flight attendant checks your boarding pass and says, "Sir, you're in economy." You look down—the pass indeed says Economy. But you paid for first class.

Even more absurdly, economy passengers found themselves in first class, while people in the waiting lounge suddenly received business class invitations.

That's what happened with ChatGPT today.

The blogger at FishAI described the chaos: friends came to him saying, "Fish, my OpenAI account was nuked overnight—I paid full price with a US card and home IP!" Meanwhile, his own account was fine—as a "professional deal hunter," he actually missed out on this "windfall," stomping his feet in frustration.

Absurd? Yes. But after the laughs, a chill runs down your spine.

Behind the Bug: Your AI Isn't Really Yours

The most unsettling thing about this Bug isn't "Pro becomes Free" itself, but the fact it exposes—your control over AI tools is far weaker than you think.

You pay monthly on time, believing you "own" Pro-level service. But in reality, your permissions, quota, and data access all live on OpenAI's servers, controlled by a system you can't see or manage. When that system bugs out, you have nowhere to seek recourse.

Think about your daily workflow: code written by ChatGPT, plans revised by ChatGPT, emails replied by ChatGPT, even meeting minutes organized by ChatGPT. If this tool suddenly drops from Pro to Free, what happens to your productivity?

This isn't just a "missing feature" issue. It's your entire work engine switching from a V8 to a tricycle.

Gray Market Chain: Another Interpretation of the Bug

Some speculate this Bug is OpenAI "cracking down on cheap Pro accounts." This claim isn't baseless.

Search results show that in April 2026, security warnings appeared in the OpenAI community—attackers using specific technical methods to bypass payment gates and bulk-grab ChatGPT Plus/Pro subscription access, known in the industry as "CDK" (subscription activation keys). The core vulnerability of the gray chain was OpenAI's trial program for Plus/Pro in the UK, in partnership with Google.

In June, V2EX and other Chinese developer communities saw a sudden wave of posts about ChatGPT and Codex accounts being banned. Some accounts were normally paid, normally used, with no traces of scripting—yet they were swept up indiscriminately.

But FishAI makes a valid point: cheap Pro accounts usually drop within a day or two. If OpenAI wanted to crack down, they could do it anytime—they wouldn't need a Bug like this. This Bug feels more like a systemic permissions management failure—a large-scale outage affecting tens of millions of users.

GPT-5.6 Leak? The Bug's Timing Is Too Coincidental

Around the same time this Bug occurred, another rumor spread through AI circles: GPT-5.6 might be undergoing stealth A/B testing.

Some ChatGPT Pro subscribers discovered that when selecting GPT-5.5 Pro, the wait time for generating complex content doubled, completely different from the usual fast output. AI influencer Leo confirmed via post that OpenAI is conducting stealth A/B testing—GPT-5.6 Pro is quietly embedded in the 5.5 Pro entry, with some paid users randomly hitting the new version.

If this claim holds, the timing of the permissions Bug is too delicate—large-scale A/B testing involves complex user grouping and permission switching; a single mistake anywhere could cause tier confusion.

Of course, this is just speculation. OpenAI hasn't provided an official explanation yet.

What You Should Worry About

Don't just spectate. This Bug should make you think about three things:

First, do you have a Plan B? If ChatGPT suddenly becomes unavailable, can you immediately switch to Claude, Gemini, or DeepSeek? Is your workflow tied to one platform or easily migrated?

Second, have you outsourced your core skills to AI? If you can't write code, design plans, or do analysis without ChatGPT, the problem isn't with OpenAI—it's with you.

Third, what exactly did you buy with your AI subscription? For $200 a month Pro, you bought a service, not ownership. Services can be interrupted, downgraded, or disappear due to bugs. You paid for "usage rights," but the word "guarantee" isn't in the contract.

Wait for an Explanation, Not Compensation

FishAI predicts OpenAI will offer affected users a free one-month equivalent membership. Based on past behavior, that's the most likely compensation.

But more than a month of Pro access, I'd like to see a detailed incident report—what exactly caused the global permissions mess? A code bug? A/B testing gone wild? Or a fundamental flaw in the permissions system architecture?

Because if even the CEO's account can fall, the reliability of this system is far lower than we think.

And the platform you bet your work, your code, and your decisions on every day—its foundation might not be as stable as you imagine.

Next time you open ChatGPT, check the tier in the bottom-left corner. Make sure it's the one you paid for.

Promotional links:

• Agnes AI: Million-context + 4K image generation + video all free → https://platform.agnes-ai.com/
• Stepfun: API free for 120 days limited → https://platform.stepfun.com?invite_code=PQCZLBNK