Claude outage on 23 October 2026
CriticalAn email when it goes down, and when it’s back.
“Elevated errors for Claude Fable 5.1, Sonnet 5.5, and Haiku 5.5”
On 23 October, Claude had a critical outage with elevated errors on requests to Claude Fable 5.1, Sonnet 5.5 and Haiku 5.5, and on fast mode requests to Opus 5.5. claude.ai, the Claude API, Claude Code and Claude Cowork were affected. The status page incident was open from 23:37 to 00:11 on 24 October, but its final update put the impact window at 23:28 to 23:59, starting before the incident was opened. The status page said a cause had been identified but did not describe it. Cursor, GitHub, Asana, OpenAI and other services also reported problems around the same time.
Claude’s own incident page ↗- When
- 23:39–00:12 UTC
- How long
- 33 min
- Incident open on its page
- 23:37–00:11 UTC
- Components affected
- 4
- Claude, 30 days before
- 3 other outages
What broke when
From Claude’s own status page: when it moved each component into trouble, and back.
At the same time
Started in the hour before or during it
And 60 lesser-known services. Problems at the same time are a clue, never proof that one caused another.
Timeline
Times in UTC
- Investigating
- claude.ai: major outage
- Claude API (api.anthropic.com): major outage
- Claude Code: major outage
- Claude Cowork: major outage
- Cause identified
- Fix in place, monitoring
- claude.ai: degraded performance
- Claude API (api.anthropic.com): degraded performance
- Claude Code: degraded performance
- Claude Cowork: degraded performance
- Resolved
- claude.ai: back to normal
- Claude API (api.anthropic.com): back to normal
- Claude Code: back to normal
- Claude Cowork: back to normal
Where this comes from
Stackpulse reads Claude’s official status page at status.claude.com every minute. The times above are when it reported the problem; the summary is written from its updates and checked against them.
Stackpulse isn’t affiliated with Claude. Names and logos belong to their owners.