Gemini status: is it down right now?
Our own check of gemini.google.com from Europe and the US, every 3 minutes, with recent history.
gemini.google.com responded. The page answered normally in 380 ms.
gemini.google.com responded.
The page answered normally in 380 ms.
Checked (1 minute ago) from Europe and the US, automatic every 3 minutes.
Run a website? Hear about it before your visitors do.
Gemini is up right now. Visual Sentinel watches your own site the same way and alerts you when a check fails or the page looks broken. Add your address and we'll check it now. No account needed.
This checks the response from gemini.google.com. It does not test playback, login or every Gemini feature.
Gemini status history
Our own checks from Europe and the US. Green means every check passed, amber means some checks failed, red means most checks in that period failed.
Outages our checks detected
None since Oct 13. A problem has to fail two checks in a row before we record it.
Watch your own site
This page is for their outage. Your own site can fail the same way. See its first check free, then try a paid plan for 7 days (card required).
Monitor my siteabout this service
About Gemini
Gemini is Google's AI assistant family, spanning the consumer app at gemini.google.com, Gemini integration inside Google Workspace (Gmail, Docs, Sheets), Gemini Advanced (paid tier), and the developer Gemini API on Vertex AI and AI Studio. Outages typically fall into consumer-side (gemini.google.com chat failing, image generation producing errors, Deep Research stuck) and API-side (Vertex AI endpoints returning 5xx, quota exhaustion even well below limits). Workspace-embedded Gemini features ("Help me write" in Docs, smart reply in Gmail) fail silently during outages, degrading gracefully back to the base product rather than showing error banners.
The official Gemini status page is available as a reference. Official pages are useful, but they can lag behind real outages by minutes or hours.
outage patterns
Gemini not working? What an outage looks like
AI assistants often fail in part rather than all at once. The site or app can open while replies stall, stop partway or end in an error, or while one model or feature is unavailable and the rest still work.
The check above requests gemini.google.com from Europe and the US every 3 minutes. It does not sign in, send a prompt or test your plan, so a passing check means the service responded, not that every model and feature works.
Common signs Gemini is having issues
- Replies stop partway, never start, or end in an error message
- One model or feature fails while the others still work
- Sign-in loops, or the app keeps asking you to log in again
- Requests fail the same way on more than one device and network
Usually not a Gemini outage
- A usage or message limit on your plan
- A company network, VPN or firewall blocking the service
- An expired session: sign out and sign in again
- A browser extension or ad blocker interfering with the page
What to do when Gemini is degraded
Try again from a second device or network, and sign out and back in. If the check above or the official status page shows a problem, it is likely on Gemini's side, so wait or switch to another tool for now. If both look fine, check your plan's limits and your network first.
monitor 24/7
Monitor Gemini automatically.
Get independent uptime alerts when the public Gemini endpoint fails a confirmed check. Solo and Starter check every 5 minutes; Business and Agency check every 60 seconds.
5-minute checks from Solo
Business and Agency support 60-second uptime checks.
Multi-region
EU & US checks. Catch regional outages.
Alerts where you work
Email and push on every plan, plus team and incident channels by plan.
free first check · 7-day trial, card required · from $15/month
why independent monitoring
Provider status references
and endpoint checks answer different questions.
Independent endpoint evidence
The regional check can show that a public endpoint is failing before or after a provider posts broader incident context. It does not validate every product subsystem.
Scheduled alerts
Monitor Gemini on your plan interval and route confirmed endpoint failures through the notification channels available to your workspace.
Multi-region verification
EU and US results can reveal a regional difference in public-endpoint reachability. Confirm the exact workflow and provider region you use.
Historical data
Track Gemini uptime over time. See trends, identify recurring issues, and make informed decisions about your dependency.
Frequently asked questions
Is Gemini down right now?
See the latest check of gemini.google.com and its timestamp at the top of this page. If Gemini is down, you can also check the official status page for extra context.
How do I check if Gemini is having issues?
This page checks the public Gemini endpoint from EU and US. You can also monitor that endpoint on a schedule and receive confirmed-failure alerts through the channels available on your plan. Test the exact provider feature you use for a complete diagnosis.
How can I get alerted when Gemini goes down?
Add gemini.google.com to Visual Sentinel and its first check is free. To keep monitoring, start a 7-day trial (card required) or subscribe now: Solo and Starter check every 5 minutes, Business and Agency every 60 seconds, with email, push and chat alerts.
Where can I find Gemini outage context?
The official Gemini status page is available as a reference. Official pages are useful, but they can lag behind real outages by minutes or hours. Visual Sentinel checks the public Gemini endpoint independently from EU and US. Paid Business and Agency plans support 60-second uptime checks.
Why does Gemini in Gmail stop suggesting replies during an outage?
Workspace-embedded Gemini features call Vertex AI asynchronously and fail open, meaning if Gemini cannot respond, Gmail silently hides the suggestion. You will notice reduced AI help without any visible error, which makes outages easy to miss.
Do Gemini outages on gemini.google.com affect the Vertex AI API?
Often yes, because both run against shared model serving capacity. But the consumer app has its own application layer that can fail independently of Vertex AI. Capacity-side incidents affect both; UI-only issues do not.
Often checked together with Gemini
When an AI assistant stops answering, people usually try a second one. Each page below has its own check, run every 3 minutes from Europe and the US.