Claude outage hits multiple Anthropic models as errors surge
Anthropic identified but had not publicly explained a multi-model Claude outage that disrupted its chatbot and coding tools. The incident adds to a string of August reliability problems.
A broad disruption with a narrow public explanation
Anthropic began investigating elevated error rates across several Claude models at 05:06 UTC on August 24, after users encountered failed responses and capacity warnings. The affected lineup included Claude Mythos 5, Fable 5, Opus 5 and Opus 4.8, while Claude.ai, Claude Code and Claude Cowork were among the services showing disruption.mashable +1
By 05:27 UTC, the company said it had identified the cause and was working on a fix, but it did not disclose that cause or provide a recovery deadline. A later update at 06:42 UTC said engineers were still trying to resolve the elevated errors.sqmagazine +1
Failures spread beyond the chatbot window
The incident mattered beyond unanswered consumer prompts because Claude is also embedded in software-development workflows. Reports collected during the disruption described interrupted Claude Code tasks, including a user who said a nine-hour process had been broken, while Downdetector showed a sharp rise in complaints around the start of Anthropic’s investigation.adigitalboom
Mashable’s own tests returned a message blaming “unexpected capacity constraints” when it tried Fable 5 and Opus 5. Its US and Southeast Asia editions also reported that rival chatbots, including OpenAI’s ChatGPT, appeared to remain available, suggesting the problem was centered on Anthropic rather than a wider generative-AI service failure.mashable +1
The extent of the damage remains unclear. Anthropic had not published request-failure rates, a geographic breakdown or separate impact figures for its chatbot, API, coding tool and Cowork product when early reports were filed.sqmagazine
A recurring reliability test
The latest disruption follows a busy month on Claude’s status history. Earlier incidents affected multiple models on August 20, Opus 5 and Haiku 4.5 on August 19, and a broader set of services on August 18; additional disruptions were recorded on August 16, 14, 7 and 6.adigitalboom
That pattern raises the stakes for developers who depend on Claude for long-running work. During an earlier August outage, more than 4,000 users reported access problems before reports fell sharply as service recovered; CNET also received confirmation from Anthropic that Claude.ai, Claude Code and the API were experiencing errors.cnet The immediate question is when normal service returns. The larger one is whether Anthropic can reduce repeat failures as Claude becomes infrastructure for work that cannot simply be restarted without cost.