Pulse
IS GROK DOWN RIGHT NOW?
Grok is not one service: grok.com, Grok in X, the iOS and Android apps and the API fail independently, so the answer to “is Grok down” depends on which surface you are on. We check the official status feed every five minutes.
The official status feed could not be reached on this page load, so we cannot confirm the current state. The incident record below is unaffected.
Outage history →Uptime by month →Latest release →Status by component →
1 process is beyond their alert threshold
Published history remains available, but new readings may be missing until collection resumes. Each process and its own threshold are shown below.
Grok status by surface
Checked
The live status feed could not be reached. The incident record below is unaffected.
Where does this status data come from?
This site is collected automatically from the official status feed. If a collector stops, it is stated here rather than left as a silent gap in the record.
Alert delivery — a process past its threshold is shown on this page and sent once by email from notify.grok46.com. One notice per incident; the alert resets after the next successful run.
Last probe · 0 failed of 96 in 24 h
Alert threshold 2 h
- Status feedclearing backlog
Last 24 h: 286 runs, seen 91590, new 83296, published 1
Alert threshold 2 h
118 incidents collected but never given a page for more than 24 hours. Publication is running and working through the backlog.
Last checked
- Changelog feedcollecting
Last 24 h: 24 runs, seen 24, new 1, published 1
Alert threshold 12 h
Last checked
- xAI documentationnothing fetched
Last 24 h: 0 runs, seen 0, new 0, published 0
Alert threshold 48 h
Nothing was fetched at all in the last 24 hours. This is a collection failure, not a quiet week.
Last checked
- Community feednot reporting
Last 24 h: 1 runs, seen 3, new 0, published 0 · 1 failed
Alert threshold 48 h · incident open since
The source was read normally and had nothing new. Rows seen counts everything the source returned; new counts what entered or changed here.
Last checked
No successful run in the last 1311 h 36 min. New readings may be missing until collection resumes.
- Weekly digestnothing fetched
Last 24 h: 0 runs, seen 0, new 0, published 0
Alert threshold 192 h
Nothing was fetched at all in the last 24 hours. This is a collection failure, not a quiet week.
Last checked
Grok outage history
Full archive →- October 31, 2026Multiple services degraded44 min
- October 29, 2026Degraded performance — Origin and Cloud Agents17 min
- October 28, 2026Multiple services degraded1 h 58 min
- October 27, 2026Elevated errors affecting repository integrations46 min
- October 26, 2026Degraded agent personalization and codebase sync3 h 17 min
- October 26, 2026Elevated errors for Claude Sonnet 5.51 h 24 min
- October 24, 2026Degraded Model Performance1 min
- October 23, 2026Investigating service outage - Grok models unavailable41 min
- October 21, 2026Elevated errors for Bitbucket Cloud integrations13 min
- October 15, 2026Elevated errors affecting Grok models6 h 56 min
- October 13, 2026Degraded performance and errors across Cursor services2 h 46 min
- October 8, 2026Elevated errors affecting Opus 5.558 min
Is Grok down, or just under high demand?
Most of the time it is the latter. Under high demand is a plan limit being reached on your account, not a service failure — the status feed can report every surface as available at the same moment. If the feed above shows every component available and Grok still refuses you, the answer is your allowance, not an outage.
How often does Grok go down?
The incident archive above is the full record since January 2025, each outage on a permanent page with its duration. For the aggregate view, uptime by month shows availability per calendar month — but only for months the archive fully covers, so no figure there is extrapolated.