What happened
Anthropic experienced a service outage affecting multiple Claude models, including claude.ai, the Claude API, Claude Code, and Claude Cowork, beginning at 07:36 UTC on July 6. The issue, which spread to Fable 5 by 08:45 UTC, caused degraded performance across four of six monitored services. Claude Console and Claude for Government remained unaffected. Anthropic resolved the issues by 09:13 UTC, having previously attributed similar outages to demand outpacing available compute capacity.
Why it matters
Unreliable access to frontier AI models introduces operational risk for platform engineers and security architects integrating these services. This outage, resolved within two hours, temporarily blocked access to critical services like Claude API and Claude Code, impacting development and deployment pipelines. Previous Anthropic outages stemmed from demand outpacing compute capacity, highlighting a recurring constraint on model availability. This follows Microsoft's temporary block of Claude Fable 5 last month, underscoring the fragility of external model dependencies.
