Anthropic Outage History

Anthropic experienced elevated error rates across Claude AI, the API at api.anthropic.com, and Claude Code beginning around 5:00 AM UTC, with the issue intensifying by 8:00 AM UTC and causing failed requests and degraded model outputs for users. The Claude status page documented the incident and listed 3 affected components: claude.ai, api.anthropic.com, and Claude Code.

Anthropic Claude 4 Opus experienced a brief degradation beginning around 12:00 PM UTC, with an elevated rate of errors affecting the Claude 4 Opus models. The Claude API endpoints api.anthropic.com, claude.ai, and Claude Code were listed as degraded on Claude's status page. The issue progressed to a degraded hint at Jun 25, 2025 3:28 PM UTC and was resolved by Jun 25, 2025 4:08 PM UTC, resulting in roughly 40 minutes of degraded performance. A user in San Ramon, California reported no response from Claude Code at 3:42 PM UTC, reflecting intermittent symptoms observed by customers. Overall, there was no full outage, just a brief window of degraded service affecting multiple Claude components.

Anthropic reported elevated errors across the Haiku family, affecting claude.ai, api.anthropic.com, and Claude Code. Users attempting to run Haiku models could have faced failed requests or degraded responses during the incident window.

The incident began at Jun 25, 2025 9:23 AM UTC and was resolved by Jun 25, 2025 10:23 AM UTC, lasting about one hour. The Anthropic status page issued 2 updates during the incident as engineers investigated and then confirmed restoration.

Anthropic Claude Opus 4 experienced degraded performance due to elevated errors. The issue affected api.anthropic.com, claude.ai, and Claude Code, potentially impacting API calls and model usage for users. The incident began around 12:05 PM UTC with an uptick in errors, and the Status page later indicated elevated errors leading to a warning at Jun 24, 2025 12:58 PM UTC.

The provider issued two status updates during the incident, and recovery was observed by Jun 24, 2025 3:08 PM UTC when the services returned to normal. The outage lasted roughly 3 hours, affecting multiple Claude Opus 4 related services.

Anthropic Claude AI experienced a brief degradation as the Claude.ai component encountered elevated tool-use errors. Users attempting to access Claude AI features were affected during the incident window. The outage began at Jun 23, 2025 6:48 AM UTC and was resolved by Jun 23, 2025 7:08 AM UTC, lasting approximately 20 minutes. The issue was confined to the Claude.ai component. The Anthropic status page posted two updates during the incident: the initial notification noting investigation into elevated tool-use errors on Claude.ai, and a follow-up confirming Claude AI was back to normal.

Anthropic Claude services, including Claude Code, claude.ai, and api.anthropic.com, experienced elevated errors and degraded performance. The incident lasted about 35 minutes, beginning at Jun 22, 2025 10:38 AM UTC and ending at Jun 22, 2025 11:13 AM UTC. Anthropic stated that an investigation was underway and later confirmed recovery as affected components returned to normal. The provider issued 2 status updates during the incident, and the issue affected 3 components.

Anthropic Haiku 3.5 experienced elevated errors that affected access to the console and API endpoints, resulting in degraded performance for a period of approximately 54 minutes. A small number of user reports surfaced around 2:45 PM UTC before the official acknowledgment, indicating issues prior to the status page update. The outage began at Jun 18, 2025 3:49 PM UTC and was resolved by Jun 18, 2025 4:43 PM UTC.

During the incident, the Anthropic status page published two updates. Initially, at Jun 18, 2025 3:49 PM UTC the service moved to a degraded state with work ongoing to fix Haiku 3.5. The affected components were console.anthropic.com and api.anthropic.com. The status then returned to normal at Jun 18, 2025 4:43 PM UTC."

Anthropic's Claude services experienced a brief degraded performance beginning at Jun 18, 2025 2:33 PM UTC when the Claude status page reported elevated errors across api.anthropic.com, Claude Code, and claude.ai. This impacted users trying to access the Claude API, use Claude Code features, or interact with the Claude.ai interface. The incident lasted approximately 1 hour and 10 minutes and was resolved by Jun 18, 2025 3:43 PM UTC. User observations began around 2:45 PM UTC.

Anthropic's Claude 4 Sonnet experienced elevated errors beginning around 7:10 AM UTC. The disruption affected three components: claude.ai, api.anthropic.com, and Claude Code, causing degraded functionality for users attempting to use Claude 4 Sonnet features. The Claude status page flagged a degraded service and began investigation, with an official update at Jun 18, 2025 8:38 AM UTC.

The incident lasted approximately 3 hours and 8 minutes, with restoration observed by Jun 18, 2025 10:18 AM UTC when the status page noted a return to normal operation. The page issued two status updates during the incident. A handful of user reports were received during the early phase, aligning with the period of degraded performance across the affected components.

Anthropic Claude experienced degraded performance with elevated errors across many models, affecting api.anthropic.com, claude.ai, and Claude Code. Users attempting to generate outputs or access Claude services during the incident may have faced failed requests, timeouts, or slower responses.

Timeline and response: The degradation began around 12:00 PM UTC and was listed as degraded by Jun 17, 2025 1:58 PM UTC. The service returned to normal by Jun 17, 2025 4:18 PM UTC after two status updates were issued by Anthropic.