Start from official sources
We begin with OpenAI's status page and help center to ground each error in what OpenAI actually documents.
Methodology
How we research ChatGPT errors
This is an independent guide, so it's worth being transparent about how it's made. We base every explanation and fix on OpenAI's own status page and help center plus reproducible troubleshooting, we separate your-side causes from OpenAI-side causes, we hedge where causes genuinely vary, and we never publish invented codes or fake incident reports. Here's exactly how we work.
Our principles
What we do — and what we don't
A short summary of the rules behind every page on this site, so you know how much weight to give what you read.
| What we do | Why | What we don't do |
|---|---|---|
| Cite OpenAI's status & help pages | They are the authoritative source for incidents and official fixes. | Treat user rumours as fact. |
| Test fixes by reproducing errors | So the troubleshooting steps actually work, not just sound plausible. | Copy fixes we haven't checked. |
| Separate your-side from OpenAI-side | It tells you whether to fix something or simply wait. | Blame your device for an outage. |
| Hedge where causes vary | The same error can have several causes. | Publish invented codes, numbers or fake incident dates. |
⚠️For anything time-specific, go to the source: open status.openai.com. We deliberately don't quote live incident details or durations here, because they change from one incident to the next — the official page always has the current picture.
Don't just sit there
Hit an error while reading this? Keep working for free.
Our advice during any error or outage is the same: don't lose the time. A multi-model workspace lets you switch to another AI and keep going, then return to ChatGPT once it recovers.
Keep working
MultipleChat AI
multiplechat.aiWhen ChatGPT errors out or hits capacity, switch to Claude, Gemini or another model in the same place and keep going for free — no waiting for OpenAI to recover. Compare answers across models too.
OfficialOpenAI Status
status.openai.comThe first place to check: live incidents, outages and maintenance.
OfficialOpenAI Help
help.openai.comOpenAI's own troubleshooting and error articles.
GuideError FAQ
chatgpterror.comEvery common ChatGPT error question answered.
How a guide gets written
Our process, step by step
Every error page on this site goes through the same process before it's published, and gets revisited when ChatGPT changes.
Reproduce the error
Where possible we recreate the error and the troubleshooting steps ourselves, so the fixes are reproducible rather than guesswork.
Classify the cause
We sort each error into your-side, rate-limit or OpenAI-side, so the advice tells you whether to fix something or wait.
Hedge where it's uncertain
When an error has several possible causes, we say so and list what to check, rather than committing to one oversimplified answer.
Avoid invented specifics
We don't publish fake incident reports, made-up error codes, or invented numbers and dates. Time-specific detail points to the live status page.
Update as ChatGPT changes
Messages, codes and fixes evolve, so we revisit the guides as the product and OpenAI's pages change.
See the principles in action: error codes · something went wrong · is ChatGPT down? · error FAQ
All error guides
ChatGPT errors: every guide
All error codes400, 401, 403, 429, 500, 502, 503, 1020.
Something went wrongThe generic error, fixed.
Network errorStops mid-reply or won't send.
Too many requests (429)Rate limits and how to avoid them.
Error code 1020Cloudflare access denied.
At capacity / 503Server overloaded.
Login problemsCan't sign in or verify.
App errorsiPhone & Android crashes.
Is ChatGPT down?Confirm an outage.
Error FAQEvery common question.
MethodologyHow this guide is researched.
FAQ
About our methodology
Short answers on our sources, why we hedge, and why we stay independent.
What sources do you use to research ChatGPT errors?
We base our explanations and fixes on OpenAI's own resources — the status page at status.openai.com and the help center at help.openai.com — combined with reproducible, hands-on troubleshooting. We treat OpenAI's official pages as the primary source and use widely observed user reports only as a secondary, cross-checking signal.
Why do your guides hedge instead of giving one definitive cause?
Because the same ChatGPT error can have more than one cause. "Something went wrong", for example, can be a server hiccup, a stuck chat, a cache or extension problem, or a network issue. We separate your-side from OpenAI-side causes and hedge where the cause genuinely varies, so you check the right things instead of trusting a single oversimplified answer.
Do you publish specific incident dates or exact error statistics?
No. We do not publish fake incident reports, invented error codes, or made-up numbers and dates. Outages and their durations vary from one incident to the next, so for anything time-specific we point you to OpenAI's live status page rather than quoting figures that would quickly become wrong or misleading.
How do you keep the guides up to date?
ChatGPT's error messages, codes and fixes change over time, so we update the guides as the product changes and as OpenAI's status and help pages are revised. Where something is likely to shift, we phrase it so it stays accurate and point you to the official pages for the current detail.
Are you affiliated with OpenAI?
No. ChatGPTError.com is an independent guide and is not affiliated with, endorsed by, or operated by OpenAI. We link to OpenAI's official status and help pages as the authoritative sources, and we always recommend verifying current details there.