OpenAI status page explained — how to read it

OpenAI's status page looks deceptively simple — green circles and timestamps — but knowing what each component actually means can save you hours of troubleshooting. The difference between "API Major Outage" and "ChatGPT

OpenAI status page explained — how to read it

OpenAI's status page looks deceptively simple — green circles and timestamps — but knowing what each component actually means can save you hours of troubleshooting. The difference between "API Major Outage" and "ChatGPT Partial Outage" tells you whether to wait five minutes or switch to a different model entirely.

How to check if OpenAI is actually down

Start with the live OpenAI status monitor on this site. It pulls directly from OpenAI's official feed and updates every 60 seconds, showing you the current state of every component without navigating multiple pages. You'll see ChatGPT, API, Sora, Playground and other services listed with operational status in real time.

For deeper detail, visit status.openai.com directly. OpenAI breaks their infrastructure into specific components: ChatGPT (web), ChatGPT (mobile), API, Labs (Sora, DALL-E), Playground, and Authentication. Each component shows one of four states: Operational (green), Degraded Performance (yellow), Partial Outage (orange), or Major Outage (red). When you see "Partial Outage" for ChatGPT but "Operational" for API, that means the chatbot interface is struggling while programmatic access still works — you could switch to using the API via Playground or a third-party interface.

The incident history section shows past outages with timestamps, duration and affected components. If you see "Identified" next to a current incident, OpenAI knows what's wrong but hasn't fixed it yet. "Monitoring" means they've deployed a fix and are watching to confirm it worked. "Resolved" means it's over. These distinctions matter — "Identified" can mean another 20 minutes of downtime, while "Monitoring" usually means you can retry in 2-3 minutes.

Is it OpenAI or is it you

Before blaming the platform, isolate whether the problem is local to you:

  • Check the live OpenAI status first. If it shows green across all components, the issue is likely on your end.
  • Open an incognito window and try chatgpt.com. If ChatGPT works there, your cookies or cache are corrupted. Go to Settings > Privacy and security > Clear browsing data, select "Cached images and files" and "Cookies" for chatgpt.com only, then try again.
  • Disable browser extensions. VPN blockers, privacy tools and ad blockers frequently break WebSocket connections that ChatGPT needs. Turn off extensions one by one or test in incognito mode (which disables most extensions automatically).
  • Try a different network. Switch from WiFi to mobile data or vice versa. Corporate and school networks often block OpenAI domains. If mobile data works but WiFi doesn't, your router or ISP is filtering the connection.
  • Test the mobile app versus browser. Download the official ChatGPT app for iOS or Android and try there. If the app works but the browser doesn't, you've got a browser-specific problem — check for ChatGPT not working — troubleshooting steps.
  • Check API error codes. If you're using the API and getting errors, look at the specific code. Error 429 means rate limiting (you're going too fast), not an outage. Error 500-series means OpenAI's servers are struggling. See OpenAI API error 429 and other API errors for the full breakdown.

What to do during an outage

When OpenAI confirms an outage, these steps minimise disruption:

  • For API users: implement exponential backoff. Don't hammer the endpoint with retries every second. Start with a 2-second delay, then 4, 8, 16, up to a 60-second maximum. This prevents you from hitting rate limits the moment service resumes.
  • Queue requests locally. Store failed requests in memory or a simple text file so you can replay them once the status page shows "Monitoring" or "Resolved."
  • Switch models if available. During partial outages, GPT-4 might be down while GPT-3.5 still works, or vice versa. Check the status page component details — if only one model is affected, specify a different one in your API calls.
  • Don't delete your account or cancel subscriptions. Outages don't justify refunds according to OpenAI's terms, and deleting your account during instability can cause permanent data loss without recovering your subscription.
  • Don't contact support about known outages. If status.openai.com shows an active incident, OpenAI's team already knows. You'll get faster resolution by waiting than by filing a ticket. For non-outage issues, see How to contact OpenAI support.

Questions people actually ask

Q: Is ChatGPT down right now?

A: Check the live OpenAI status on this site — it updates every 60 seconds and shows you the current state of ChatGPT web, mobile, API and all other services immediately.

Q: What does "Degraded Performance" actually mean?

A: The service is running but slower than usual. You might see longer response times, occasional timeouts, or features like image generation taking 30 seconds instead of 10.

Q: Why does the status page say "Operational" when ChatGPT isn't working for me?

A: Operational means OpenAI's core infrastructure is healthy, but you could still have network issues, account problems, browser conflicts or regional routing failures. Follow the isolation steps above.

Q: How long do outages typically last?

A: Minor incidents resolve in 10-30 minutes. Major outages can take 2-4 hours. Check the incident updates on status.openai.com — they usually post time estimates once they've identified the root cause.

What to remember

  • The live OpenAI status updates every 60 seconds from OpenAI's official feed — bookmark it
  • Component-level status tells you what's affected — API down doesn't mean ChatGPT is down
  • "Identified" means they know the problem, "Monitoring" means the fix is deployed, "Resolved" means it's over
  • Isolate local issues before assuming OpenAI is down — incognito mode catches 80% of false alarms
  • During outages, implement exponential backoff for API calls and never spam support about known incidents

---

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

Related help