AI STATUS BOARD

Claude API outage report · Tuesday 22 September 2026

Claude API was
down for
1 h 38 min.

Anthropic reported “Elevated errors for multiple models”. 1 h 38 min in total, resolved at 02:35 UTC.

In numbers

RESOLVED 02:35 UTC
Down in total
1 h 38 min
Outages
1
Longest
1 h 38 min
Updates from Anthropic
5
Parts hit
  • claude.ai
  • Claude API (api.anthropic.com)
  • Claude Code
  • Claude Cowork

The day, hour by hour

UTC
00:0006:0012:0018:0024:00

What Anthropic said, as it happened

00:57 → 02:35 UTC · Elevated errors for multiple models

1 h 38 min
  1. 02:35resolved

    This issue has been resolved. Impact occurred from 5:50pm PT / 00:50 UTC to 7:10pm PT / 02:10 UTC.

  2. 02:11monitoring

    We have seen success rates return to normal across affected models, and are monitoring closely to ensure no further issues.

  3. 01:35identified

    We are continuing to work to resolve errors affecting some models. At this time, requests to Claude Fable 5 and 5.1 and Mythos 5 and 5.1 have returned to normal success rates. We are working to resolve remaining errors affecting Claude Opus 5, and will provide an additional update shortly.

  4. 01:17identified

    We have identified the cause of elevated errors on requests to Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 and are working on a fix. We will provide an update as soon as possible.

  5. 00:57investigating

    We are investigating elevated errors on requests to Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5. We will provide an update as soon as possible.

Anthropic's incident page

Quoted from Anthropic's status page. Times in UTC.

Was it just Claude API?

No. Replit also had problems at the same time.

  • Replit00:22 (18 Sep) → 14:52 (7 Oct)GitHub connector may not be working for users

Other problems that day:

  • Mistral00:42 → 01:14Mistral OCR 4 is degraded
  • Lovable07:53 → 08:07Elevated errors when Enabling Lovable Cloud
  • Mistral09:16 → 09:53Skill creation and updates unavailable
  • Cursor13:02 → 13:45Investigating service degradation — Automations and Cloud Agents w/ Grok 4.7
  • Replit21:36 → 23:42Connector status mismatch and removal errors

Claude API's record

Incidents, last 90 days
31
Typical length
1 h 2 min

This day's longest outage ran 1 times the typical length.

All Claude API outages

Hear first next time

A notification the moment Claude API breaks, and when it's back.

Questions about this outage

Why was Claude API down on 22 September 2026?

Anthropic described it as “Elevated errors for multiple models”. Its updates are quoted on this page; if it posts a cause later, it shows up there.

Which parts of Claude API were affected?

Anthropic listed: claude.ai, Claude API (api.anthropic.com), Claude Code, Claude Cowork.

How long was Claude API down?

1 h 38 min, from 00:57 UTC to 02:35 UTC.

Were other AI tools down at the same time?

Yes: Replit reported problems during it.

How do I know next time Claude API goes down?

Turn on alerts on AI Status Board. You get a notification the moment Claude API breaks and again when it is back.