Claude Outage History
Every incident Anthropic has published for Claude recently, newest first, with how long each one lasted.
Showing 40 published incidents back to 26 Jul 2026.
- Major15 Sep 20261h 26m
Intermittent error spikes for Claude Mythos 5.1 and Claude Fable 5.1
- Major11 Sep 202617 min
Elevated errors for Claude Mythos 5.1 and Claude Fable 5.1
- Minor10 Sep 20261h 41m
Elevated latency on the Claude API for some users
- Major10 Sep 20264d 3h
Degraded functionality for Claude Cowork on Windows
- Major3 Sep 20262h 57m
Elevated errors for multiple models
- Major3 Sep 202619 min
Elevated errors for Claude Sonnet 5
- Major2 Sep 202627 min
Elevated errors for Claude Sonnet 5
- Minor1 Sep 20261h 58m
Delays in credit purchases
- Minor1 Sep 20261h 2m
Degraded performance on platform.claude.com and Claude for Microsoft Office 365
- Minor1 Sep 202620 min
Degraded performance on Claude for Microsoft Office 365
- Minor31 Aug 20261h 14m
Degraded performance on claude.ai and Claude Code
- Minor31 Aug 202629 min
Degraded performance on claude.ai
- Major28 Aug 20262h 59m
Elevated errors on Claude Code and Claude Cowork
- Minor24 Aug 202615 min
Issues logging into Claude.ai
- Minor24 Aug 20263 min
Errors logging into Claude.ai
- Major24 Aug 20263h 24m
Elevated errors for multiple models
- Major20 Aug 202626 min
Elevated errors on requests to multiple models
- Minor20 Aug 202629 min
Elevated errors on Google connectors
- Minor19 Aug 20261h 20m
Degraded performance for Claude Opus 5 and Claude Haiku 4.5
- Minor18 Aug 20262h 41m
Degraded performance for multiple models
- Minor17 Aug 20261h 33m
Degraded performance for Claude Opus 5, Claude Sonnet 5
- Critical16 Aug 202636 min
Service disruption on Claude services
- Major14 Aug 202630 min
Elevated errors for Claude Fable 5
- Major14 Aug 20261h 28m
Service disruption on Claude services
- Minor14 Aug 202615 min
Service disruption on Claude Code on the web and related services
- Major13 Aug 20261h 35m
Elevated errors for Claude Mythos 5, Claude Fable 5, and Claude Sonnet 5
- Minor12 Aug 20264h 17m
Degraded performance for multiple models
- Minor5 Aug 202643 min
Degraded performance for Claude Opus 5
- Minor5 Aug 20267h 9m
Degraded performance of multiple models
- Minor4 Aug 202625 min
Elevated errors on Claude Sonnet 5
- Minor4 Aug 202639 min
Elevated errors on Claude Sonnet 5
- Minor3 Aug 202615 min
Degraded performance on Claude Sonnet 5
- Minor3 Aug 20261h 24m
Error rates across multiple models
- Minor31 Jul 202647 min
Degraded performance on Claude Sonnet 5
- Minor30 Jul 202641 min
Degraded performance on Claude Opus 4.8
- Major30 Jul 20264h 51m
Elevated errors across many models
- Minor27 Jul 202657 min
Elevated errors on Claude Opus 5 and Haiku 4.5
- Minor27 Jul 20261h 3m
Elevated errors on Claude Opus 5
- Minor27 Jul 202648 min
Elevated errors on Claude Opus 5
- Major26 Jul 20261h 27m
Elevated errors for Opus 5
What this record shows
Anthropic reports at model granularity, so incident titles here often name a specific model rather than the whole service. An entry about one model usually means the others kept working, which is why "Claude is down" is worth narrowing before you conclude the service failed.
Claude reaches users through several surfaces — the web app, the API, Claude Code and the developer console — and Anthropic reports on each separately. An incident confined to the API can leave the chat interface untouched, and the reverse happens too.
Major and critical entries are the ones that cost real availability. Minor entries are typically brief periods of elevated errors where most requests still succeeded.
How to read the durations
Each entry runs from when Anthropic opened the incident to when it was marked resolved. That is a generous measure: an incident is usually opened once the problem is confirmed and closed after it has been stable for a while, so the window is wider than the period in which most people were actually affected.
Severity comes from the provider. Only major and critical entries count against the uptime figure on the Claude status page, because a minor incident normally means degraded performance for some users rather than the service being unavailable.
This list is limited to what the provider has published and still exposes. Providers cap how far their public feed reaches, so an empty stretch further back means the record has aged out, not that nothing happened.
🔗 Keep checking