Anthropic Outage History

Anthropic Claude experienced a temporary service disruption beginning at 9:23 PM UTC that affected Claude.ai, platform.claude.com, and Claude Code. Users reported being logged out and unable to sign back in, including observations from Brisbane, Australia. The incident manifested as degraded authentication and login-related issues across Claude services, impacting a broad but not universal set of users.

Anthropic began investigating immediately and released two status updates during the incident. The outage persisted for roughly 3 hours and 20 minutes, ending when services were brought back up at 12:43 AM UTC. Report activity showed a spike early in the event, then gradually subsided as systems recovered. The issues were contained to authentication and service access rather than a full platform-wide failure.

Anthropic Claude experienced a global login outage that left users unable to sign in and the Claude app unavailable for a brief period. The outage began around 9:00 PM UTC and was resolved by 9:18 PM UTC. StatusGator's Early Warning Signal detected login issues at Oct 9, 2025 9:05 PM UTC, about 13 minutes before the incident ended, and before Anthropic published any status updates. User reports spanning North America, Europe, and Australia described sign-in failures, sign-in errors, or app errors, indicating a global impact on authentication and access to Claude.

Anthropic reported elevated errors affecting Claude Sonnet 4.5 across four components: platform.claude.com, Claude API, claude.ai, and Claude Code. Users experienced slow responses and intermittent failures while interacting with Claude Sonnet 4.5. The incident began at Oct 8, 2025 1:23 PM UTC and was resolved by Oct 8, 2025 2:28 PM UTC, lasting about 1 hour and 5 minutes. The Claude status page listed the issue as Investigating with elevated errors and then returned to normal after restoration, with two official updates issued during the incident.

An outage briefly degraded Claude Opus 4 services, affecting the Claude API at api.anthropic.com, claude.ai, and platform.claude.com. The issue began on Oct 6, 2025 8:33 AM UTC and lasted about 10 minutes, with users potentially experiencing elevated errors or slower responses when using the API, Claude AI, or the Claude console. The Claude status page noted that the issue had been identified and a fix was being implemented, and services returned to normal after the subsequent update. The incident involved two official status updates during its course.

Anthropic Claude experienced a global outage with widespread 503 errors, upstream connect failures, and degraded performance across the Claude API, UI, and related tooling. Users across North America, Europe, and Asia-Pacific reported API errors, timeouts, and slow responses, affecting access to Claude features and integrations such as IDE plugins.

StatusGator logged an Early Warning Signal starting at 1:21 AM UTC and ending at 3:07 AM UTC due to rate limit errors and 500 server errors on Claude service pages. The provider’s status page did not show updates in this feed during the period, while user reports remained high across regions.

The incident demonstrates global reach with sustained issues through the early-morning hours, based on public-facing reports across multiple regions.

Anthropic Claude Connectors experienced a brief connectivity outage affecting Claude.ai Connectors / MCP, preventing users from connecting to Claude.ai Connectors. The disruption lasted about 45 minutes, beginning around 4:48 PM UTC and ending by 5:33 PM UTC. During the incident the Claude.ai component was listed as a degraded status (warn) on the Claude status page.

The provider issued 2 status updates over the incident: the first at 4:48 PM UTC noting Claude.ai Connectors are Unhealthy and that investigators were looking into the issue, and the second at 5:33 PM UTC confirming restoration to normal as the status returned to up. The issue affected one component (Claude.ai).

Anthropic Claude experienced elevated errors resulting in degraded performance for Claude AI and the platform.claude.com experience. Users reported slower responses and occasional failures, including a report from Denver, Colorado. The incident began at Sep 30, 2025 6:58 PM UTC and was resolved by Oct 1, 2025 12:08 AM UTC. The status page noted an investigation with elevated errors, and two official updates were posted as the services moved from the degraded state back to normal for both components.

Anthropic reported a brief degradation across Claude Sonnet 4, impacting Claude API (api.anthropic.com), claude.ai, and platform.claude.com. The issue began at 10:43 AM UTC and prompted an investigation by the Claude team. The services returned to normal by 1:38 PM UTC, after roughly 2 hours and 55 minutes. The provider issued 2 status updates during the incident as it worked to restore full functionality across the Claude stack (API, Claude AI, and platform).

Anthropic reported elevated errors affecting Claude API (api.anthropic.com), claude.ai, and platform.claude.com. The issue began at Sep 30, 2025 10:03 AM UTC and was resolved by Sep 30, 2025 10:18 AM UTC, lasting about 15 minutes. Users may have experienced errors when using Claude Sonnet 4 features during this window. The incident impacted three components and was followed by two status updates from Anthropic as the team worked to restore services.

Two official updates were published during the incident, indicating the elevated errors and subsequent recovery across the affected components. Services returned to normal after Sep 30, 2025 10:18 AM UTC according to the status page.

Anthropic Claude Sonnet 4 experienced elevated errors impacting Claude API (api.anthropic.com), claude.ai, and platform.claude.com. Users attempting to use these services likely encountered errors or degraded responses. The incident began at 7:38 AM UTC when the Claude status page moved to warning and indicated an investigation, and it lasted for about 45 minutes, with services returning to normal by 8:23 AM UTC. The provider issued two status updates during the incident.