Claude's outage: A recent stumble for Anthropic

      Update: Anthropic has reported that the issue has been identified and a solution is being put in place. The company states that the problem affects Claude.ai, Claude Console, Claude API, Claude Code, and Claude Cowork.

      Anthropic has found a fix for the increased error rates impacting various Claude models, even as it continues to explain the reasons behind the suspension of access to Claude Mythos 5 and Claude Fable 5. According to Anthropic's status page, the company began examining the elevated error rates across several models on June 23 at 14:19 UTC and went on to identify the issue by 14:25 UTC, with a resolution being applied. While this may sound orderly in theory, in practice, it occurs during a month that has already tested users' tolerance for operational instability from one of the most prominent AI providers.

      The current incident is only a part of the overall situation. On June 13, Anthropic announced the suspension of access to Claude Mythos 5 and Claude Fable 5, impacting Claude.ai, Claude API, Claude Code, and Claude Cowork. At that time, the company indicated that the models were under monitoring and directed readers to its announcement for further details.

      The latest updates from the EU tech scene, a story from founder Boris, and some questionable AI art are available for free every week in your inbox. Sign up now! These two events now fit within a larger trend, as Anthropic has recorded several recent status updates regarding increased errors that have impacted other Claude models, suggesting that the issues from June are more than just isolated outages.

      Discussions in the ClaudeAI community on Reddit reflect this sentiment, with users reporting outages, retries, and the increasingly tedious task of adapting to provider instability rather than avoiding it.

      What the status page conveys

      The company's timeline is notably clear. It first categorized the incident as under investigation, then identified, all within six minutes on June 23. This indicates a swift operational response, yet it does not inform users about the extent of the impact or the duration of the degraded performance.

      Anthropic has not provided a root-cause analysis for this recent incident in the publicly available material examined here. This lack of information is significant. In a market where models are increasingly seen as foundational infrastructure, distinguishing between a temporary issue and a persistent reliability concern is crucial for commercial interests.

      A month of repetition

      The suspension of Claude Mythos 5 and Claude Fable 5 is the more significant issue since it raises questions that extend beyond system uptime. Anthropic has indicated that the matter was serious enough to completely revoke access, which sends a stronger message than the typical “we're investigating” language commonly found on status pages.

      As a result, the company must reassure users on two fronts simultaneously: that the models function correctly, and that they can continue to operate. This is why the current event appears more significant than its timestamp suggests. Even if the latest error surge resolves quickly, it will still be viewed in the context of the previous suspension in June. If another outage occurs, the pattern may no longer seem random but rather a sign of product maturity colliding with real-world dependency.

      The pressing question is not whether Claude will recover from a single unfortunate incident, but whether Anthropic can transform a month of status-page fluctuations into something more enduring than mere reassurance.