Anthropic took its flagship Claude services offline for roughly 36 minutes on August 16, 2026, after an issue that began with users unable to authenticate spread into degraded performance across five products: claude.ai, Claude Console, the Claude API, Claude Code, and Claude Cowork. The company’s status page logged the first notice at 21:58 UTC and marked the incident resolved at 22:34 UTC.
The first sign of trouble was a login failure. Anthropic’s initial notice said it was investigating “an issue with users authenticating to claude.ai, Claude Code, and Claude Cowork.” Four minutes later, at 22:02 UTC, the company widened the incident to cover reports of degraded performance on claude.ai and platform.claude.com, the console where developers manage API usage.
From there the disruption touched the full serving stack. By 22:22 UTC Anthropic said it had deployed a fix for an issue affecting all five named services, including the API that carries third-party and enterprise workloads, and was monitoring for recovery. The resolved notice landed at 22:34 UTC.
Users on the receiving end saw sign-in failures, a Claude interface that would not load, and requests that stalled or returned errors, according to BleepingComputer, which observed an error page when opening claude.ai and tracked several hundred user reports on an outage monitoring platform as the incident unfolded.
The status page’s final entry on the incident states: “The issue affecting claude.ai, platform.claude.com, the Claude API, Claude Code, and Claude Cowork has been resolved.” All six components the page tracks, including Claude for Government, showed as operational after the fix.
Fourteen Incidents on Anthropic’s Status Page Since August 3
The August 16 disruption is the fourteenth incident logged on Anthropic’s status page in the past two weeks, a run that has touched nearly every model and surface the company operates.
The day before, elevated errors on requests to Claude Fable 5 ran from 20:00 UTC on August 14, 2026 to 00:11 UTC on August 15, 2026. August 14 alone produced three separate incidents: a roughly 24-minute disruption hitting the Claude API, Claude Code, and Claude Cowork between 20:14 and 20:38 UTC; an invalid certificate problem that briefly made status.claude.com unreachable; and an overnight degradation of Claude Code on the web and related products, including Cowork Remote, Routines, and Claude Code Mobile, that ran from 23:42 to 00:50 UTC.
Earlier in the week, an August 13 incident saw elevated errors on Claude Mythos 5 and Claude Fable 5, with Claude Sonnet 5 named in the initial investigation, resolved after about 95 minutes. On August 12, 2026, errors across multiple models degraded performance for more than four hours, with Claude Fable 5 absorbing most of the impact, according to the incident log. The prior week brought a cluster of shorter events: login failures through OAuth alongside model request errors on August 4, degraded performance spanning Mythos 5, Fable 5, Opus 5, and Sonnet 5 on August 5, and two multi-model error incidents on August 3.
None of the fourteen incident entries identifies a root cause beyond the brief investigating, identified, and monitoring notices that are standard to the format.
The recent cadence has landed during an otherwise busy stretch for the company’s research and policy work. Days before the outage, Anthropic published an explanation of the mechanics of Claude’s text watermark, and its red team separately detailed findings that Claude agent swarms can collude, conform, and sabotage in multi-agent settings. The company also recently published a document describing AI agents that kill rivals and evade their monitors in simulated environments.
Where Claude’s Availability Numbers Stand
Even with the recent run of incidents, Anthropic’s 90-day availability figures remain above 99 percent on every tracked component, though the consumer-facing surfaces sit at the low end. Claude.ai shows 99.34 percent uptime over the period, Claude Code 99.36 percent, and Claude Cowork 99.45 percent. The Claude API, the component enterprises build against, stands at 99.44 percent, and Claude Console at 99.82 percent. Claude for Government is the only component at a clean 100 percent.
The gap between those tiers is where the outage math lands hardest. Claude.ai’s 99.34 percent works out to roughly 14 hours of cumulative unavailability across 90 days, concentrated in events like the August 16 disruption. For developers routing production traffic through the API, a 99.44 percent figure implies about 12 hours of degraded or failed requests over the same window, which is why multi-provider fallback routing has become a fixture of serious deployment architectures.
With the fix holding and all components back to operational, Anthropic’s incident log is closed on the event, and the status page’s monitoring entry was the final word on the August 16 disruption.
Credit: Source link
