Claude Was Down 7.5 Hours Today — What It Means for Engineers Building on the API

All four Claude frontier models — Mythos 5, Fable 5, Opus 5, Sonnet 5 — went down simultaneously for 7.5 hours today; this was Anthropic's 164th documented outage since January.

The incident started at 3:05 AM ET when elevated errors hit all four models at once. Engineers pushed an initial fix at 9:08 AM ET, then a second Opus 5-specific degradation emerged at 9:51 AM, requiring its own resolution path. Both were fully resolved by 10:34 AM ET — a combined 7 hours 29 minutes of disruption.

According to StatusGator, Anthropic has experienced 164 outages in 2026 — roughly one incident every 1.3 days. The platform's 90-day uptime figures translate to 14–23 hours of downtime per service per quarter, well short of the 99.9% enterprise standard.

Why it matters: Any agentic pipeline that calls a single provider with no fallback inherits that provider's downtime. Today's outage is a good prompt to add a model fallback chain — OpenRouter makes this a one-field change in your API call (see today's tutorial).

Sources: Claude outage — Android Authority · Outage history — StatusGator · Benzinga