What to do during a ChatGPT outage

You're staring at a spinning wheel or a connection error, and you need to know: is ChatGPT actually down or is this a problem on your end? Check the [live OpenAI status](/status) on this site first — it updates every 60

What to do during a ChatGPT outage

You're staring at a spinning wheel or a connection error, and you need to know: is ChatGPT actually down or is this a problem on your end? Check the live OpenAI status on this site first — it updates every 60 seconds from OpenAI's official feed and shows you exactly which services are affected.

How to check if OpenAI is actually down

Start with this site's live OpenAI status monitor. It pulls data directly from status.openai.com every minute and breaks down the health of individual components: ChatGPT web interface, API servers, Sora, and plugins. If you see "Operational" in green next to ChatGPT, the problem is likely on your side.

Head to status.openai.com as your second source. OpenAI posts incident updates here within minutes of detecting widespread issues. Look for orange or red indicators and click into individual components to see granular status. During partial outages, you might see "ChatGPT" marked operational but "API" degraded — which matters if you're using third-party apps that rely on OpenAI's backend.

Check third-party reports on Twitter/X by searching "ChatGPT down" or "OpenAI down" and filtering by "Latest". Downdetector shows user-submitted reports in real time. If thousands of people started reporting issues in the last 10 minutes and your status checks confirm problems, it's an outage.

Is it OpenAI or is it you

Before you assume it's an outage, run these isolation tests. They take three minutes and solve 70% of "ChatGPT won't work" problems that aren't actual outages.

  • Open an incognito/private window and go to chatgpt.com. Sign in fresh. If ChatGPT loads normally here, the issue is browser cache, cookies, or an extension interfering with your regular session.
  • Clear site data for chatgpt.com specifically. In Chrome: Settings → Privacy and security → Site settings → View permissions and data stored across sites → search "chatgpt.com" → Clear data. In Firefox: Options → Privacy & Security → Cookies and Site Data → Manage Data → search "openai.com" → Remove Selected. This doesn't log you out of other sites.
  • Disable all browser extensions temporarily. Ad blockers, privacy tools, and VPN extensions routinely break ChatGPT's WebSocket connections. Turn them off, reload chatgpt.com, and test. If it works, re-enable extensions one by one to find the culprit.
  • Try a different network. Switch from WiFi to mobile data or use a different WiFi network. Corporate networks, university systems, and some ISPs block WebSocket protocols ChatGPT requires. If switching networks fixes it, the block is network-level.
  • Test the mobile app vs browser. If the iOS or Android app works but the website doesn't (or vice versa), the outage is component-specific or you've got a browser-specific issue. See ChatGPT not working — troubleshooting for detailed browser fixes.

What to do during an outage

When you've confirmed it's a genuine OpenAI outage, take these specific actions depending on how you use ChatGPT.

For web interface users:

  • Don't refresh obsessively. Hammering the refresh button during an outage delays recovery for everyone. Check the live OpenAI status every 15 minutes instead.
  • Don't delete your account or cancel Plus. People panic during outages and cancel subscriptions. You won't get a prorated refund for downtime — OpenAI's Terms of Service cover this explicitly.
  • Save your work elsewhere. If you're mid-conversation when an outage hits, copy critical responses to a text file immediately. ChatGPT conversation history can fail to sync during degraded performance.

For API users:

  • Implement exponential backoff immediately. Start with 1-second delays, double after each failed request: 1s, 2s, 4s, 8s, up to 60s max. This is standard practice and what OpenAI's API documentation recommends.
  • Queue requests instead of dropping them. Store failed API calls in a local queue and retry them when status.openai.com shows recovery. Don't spam the API with immediate retries — you'll hit rate limits the moment service resumes.
  • Monitor error codes. A 503 Service Unavailable means OpenAI's infrastructure is overloaded. Error 429 during an outage means you're hitting rate limits — see OpenAI API error 429 and other API errors for handling strategies.

Don't contact support during confirmed outages. OpenAI support can't restore service faster or offer individual updates. If you need human help after an outage is resolved, see How to contact OpenAI support for the actual channels that work.

Questions people actually ask

Q: Is ChatGPT down right now?

A: Check the live OpenAI status on this page — it updates every 60 seconds with real data from OpenAI's status feed. Green means operational, yellow means degraded, red means outage.

Q: Will I get refunded for ChatGPT Plus downtime?

A: No. OpenAI's Terms of Service don't guarantee uptime SLAs for ChatGPT Plus. Enterprise API customers have separate contractual agreements that may include credits.

Q: How long do ChatGPT outages usually last?

A: Most resolve in 15-45 minutes. Major incidents affecting multiple regions have lasted 2-4 hours historically. Check status.openai.com for incident-specific updates.

What to remember

  • Check this site's live status monitor and status.openai.com before assuming it's an outage
  • Run isolation tests in incognito mode, clear chatgpt.com cookies, and disable extensions first
  • Never delete your account, cancel Plus, or spam refresh during confirmed outages
  • API users must implement exponential backoff and queue failed requests
  • OpenAI support cannot accelerate recovery during infrastructure outages

---

*openai-support.com is an independent resource, not affiliated with OpenAI Inc.*

Related help