Claude users woke up to error messages again. Anthropic confirmed a fresh outage early on September 22, 2026, and this time it hit some of the company’s newest models. The issue has since been fixed, but it’s the latest in a string of disruptions that have hit Claude’s infrastructure over the past few months.
What Happened?
Anthropic’s status page flagged the problem at 12:57 AM UTC on September 22. The company said it was investigating elevated errors on requests to Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5. Downdetector reports backed that up fast, with more than 1,600 complaints piling in from users across multiple regions.
Reports weren’t limited to the chat interface either. A number of developers flagged problems with Claude Code, the tool many use for AI-assisted coding and automated workflows. When Claude Code goes down mid-session, it can interrupt builds, pipelines, and anything else running on top of the API.
Key Details
Anthropic moved fast once the issue was flagged. Here’s roughly how the timeline played out:
- 12:57 AM UTC: Anthropic starts investigating elevated errors on Mythos 5.1, Fable 5.1, and Opus 5
- 1:17 AM UTC: The company says it has found the cause and is working on a fix
- 1:35 AM UTC: Fable 5, Fable 5.1, Mythos 5, and Mythos 5.1 return to normal success rates
- 2:11 AM UTC: The incident moves to “monitoring,” with all affected models reporting as fixed
Anthropic hasn’t shared a public root cause for this specific incident. Claude.ai, Claude Code, the API, and Console are all listed as operational again on the official status page.
A Pattern of Repeated Outages
This isn’t a one-off. Claude has now logged repeated outages through the second half of 2026, including a separate incident on September 3 that also hit Mythos 5.1, Fable 5.1, and Opus 5. Anthropic launched Fable 5.1 and Mythos 5.1 on September 1, calling them its capable models for coding and knowledge work.
For casual users, an outage means a few frustrating minutes of failed replies. For businesses and developers who’ve built products on top of Claude’s API or lean on Claude Code for daily work, it’s a bigger deal. A dropped request mid-task can stall an automated workflow or interrupt a coding session with no warning.
As Anthropic competes harder for enterprise customers, one question remains: can its infrastructure keep up with rapidly growing demand? Rapid Claude Code adoption and a surge in new signups have contributed to earlier outages this year.
Read Also: OpenAI Launches GPT-6 Astra, Says We’re Now in the “AGI Era”
Official Statements
Anthropic has not issued a detailed public explanation for the September 22 incident beyond its status page updates. The company’s official status page remains the most reliable place to check real-time service health at status.claude.com.
Industry Impact
Recurring outages put pressure on Anthropic to shore up reliability even as it races rivals like OpenAI and Google on model capability. Some businesses have started building in fallback options, such as routing requests to a second AI provider when Claude’s API goes down, to avoid losing functionality during incidents like this one.
Should You Be Worried?
Today’s outage was resolved in roughly 75 minutes after being flagged, faster than several of Anthropic’s earlier incidents this year. Still, the frequency of these disruptions is becoming a talking point of its own. If you rely on Claude staying online, monitor its status page and keep a backup plan ready for unexpected outages.
