Update: Anthropic says the incident has been identified and a fix is being implemented. The company says the issue affects Claude.ai, Claude Console, Claude API, Claude Code, and Claude Cowork.
Anthropic says it has identified a fix for elevated errors across multiple Claude models, just as the company is still explaining why it suspended access to Claude Mythos 5 and Claude Fable 5.
Anthropic’s status page says the company began investigating elevated error rates across multiple models at 14:19 UTC on 23 June and moved to identified by 14:25 UTC, with a fix being implemented. That sounds tidy on paper. In practice, it lands in the middle of a month that has already tested how much operational wobble users will tolerate from one of the most visible AI providers.
The immediate incident is only part of the story. On 13 June, Anthropic posted that it had suspended access to Claude Mythos 5 and Claude Fable 5, affecting Claude.ai, Claude API, Claude Code, and Claude Cowork. The company said at the time that the models were under monitoring, and pointed readers to its announcement for more detail.
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!
Those two events now sit on top of a broader pattern. Anthropic has logged several recent status updates for elevated errors, including incidents affecting other Claude models, which gives the June wobble a less comforting shape than a one-off outage.
Reddit threads in the ClaudeAI community mirror that mood, with users describing outages, retries, and the increasingly unromantic work of building around provider instability rather than assuming away from it.
What the status page says
The company’s own timeline is unusually crisp. It first marked the incident as investigating, then as identified, all within six minutes on 23 June. That suggests an operational team moving fast, but it does not tell users how widespread the impact was or how long degraded performance lasted.
Anthropic has not, in the public material reviewed here, published a root-cause analysis for the latest incident. That omission matters. In a market where models are increasingly treated like infrastructure, the difference between a temporary fault and a recurring reliability problem is not semantic, it is commercial.
A month that keeps repeating itself
The suspension of Claude Mythos 5 and Claude Fable 5 is the more consequential thread, because it raises questions that go beyond uptime. Anthropic has said the issue was serious enough to pull access entirely, which is a stronger signal than the usual “we’re investigating” language that fills most status pages.
The result is a company that has to reassure users on two fronts at once: that the models work, and that the models are allowed to keep working.
That is why the current incident feels bigger than its timestamp. If the latest error spike clears quickly, it will still be read through the lens of June’s earlier suspension. And if another outage follows, the pattern stops looking accidental and starts looking like product maturity meeting the brick wall of real-world dependence.
The more important question now is not whether Claude recovers from one bad afternoon. It is whether Anthropic can turn a month of status-page churn into something more durable than reassurance.
Get the TNW newsletter
Get the most important tech news in your inbox each week.