Anthropic Outage History

Anthropic reported elevated errors on Claude Sonnet 3.7, impacting four components: platform.claude.com, claude.ai, Claude API (api.anthropic.com), and Claude Code. Users attempting to use Claude Sonnet 3.7 features may have experienced errors or degraded performance. The incident began at Jan 28, 2026 10:18 AM UTC and lasted for about 20 minutes, until Jan 28, 2026 10:38 AM UTC when services returned to normal.

The Claude status page published two updates during the incident, starting with an investigation notice and followed by confirmation that the issue had been resolved as services came back online. The outage was brief and affected four components, with all services restored by Jan 28, 2026 10:38 AM UTC.

Anthropic Claude Console (platform.claude.com) experienced a brief degraded performance event, impacting users trying to access or manage Claude Console tasks. The incident began around 10:28 PM UTC with an official note that the team was investigating, and the status was raised to degraded at 10:33 PM UTC. Users reported slower responses and intermittent access during the window.

Anthropic Claude Haiku 3.5 experienced elevated errors across Claude services for roughly 30 minutes on 2026-01-27. Four components were affected: platform.claude.com, claude.ai, Claude API, and Claude Code, likely resulting in degraded interactions for users. The Claude status page listed a degraded performance during the incident, with two updates at 2:28 PM UTC and 2:58 PM UTC and services returning to normal by 2:58 PM UTC. User reports were limited, indicating the impact was not globally pervasive.

Anthropic experienced an outage affecting Opus 4.5 with an increased rate of errors across Claude Code, platform.claude.com, claude.ai, and the Claude API. Users attempting to use Claude services during the incident may have encountered errors or degraded responses. The issue began at Jan 25, 2026 8:08 AM UTC with the status page noting investigation for Opus 4.5. The outage persisted until Jan 25, 2026 9:58 AM UTC when services returned to normal.

The Claude API (api.anthropic.com) experienced a brief degraded performance beginning at 5:13 AM UTC after elevated errors pointing to an invalid signature in the thinking block. Users attempting to use Claude API features that rely on the thinking block may have encountered failures or slower responses during the outage. The issue was resolved by 7:48 AM UTC, with service returning to normal as error rates stabilized.

The provider issued two status updates during the incident, noting that a fix was in progress and that the Claude API would be restored once the issue was resolved.

Anthropic Claude AI experienced a brief degraded performance incident on 2026-01-22, impacting claude.ai and Claude Code. The issue began around Jan 22, 2026 4:06 PM UTC and was first listed as degraded at Jan 22, 2026 4:08 PM UTC as user reports increased with errors, including OAuth failures and 500 responses. The outage lasted about 47 minutes, intermittently affecting user workflows across the Claude AI platform.

Anthropic provided three official updates during the incident, with claude.ai and Claude Code moving from degraded to up by Jan 22, 2026 4:53 PM UTC. There was a sizable volume of user reports reflecting the disruption during the event.

Anthropic Claude Sonnet 4.5 experienced elevated errors that affected multiple Claude services. The incident impacted users attempting to access Claude Sonnet 4.5 features through platform.claude.com, claude.ai, the Claude API, and Claude Code, with some users seeing errors or degraded performance. According to Anthropic's status updates, investigation began around Jan 21, 2026 2:44 PM UTC and the services were restored by 3:33 PM UTC, for a duration of approximately 45 minutes. The Claude status page published two updates during this window. User reports were limited, suggesting a relatively small number of affected users, with issues resolved once services returned to normal.

The outage affected Claude Opus 4.5 and Sonnet 4.5 beginning at 4:43 PM UTC, with elevated errors across Claude platform services. The issue impacted four components: platform.claude.com, claude.ai, Claude API (api.anthropic.com), and Claude Code, causing user-visible problems such as requests not responding, timeouts, and overload messages.

The Claude status page documented an investigation and later restoration. The incident was resolved by 5:38 PM UTC after roughly 55 minutes, and two status updates were posted during the course of the incident. User reports described overload and retry attempts as the outage persisted.

Anthropic's Claude AI component claude.ai experienced compaction-related issues that led to degraded performance for some users. Reports described failures to load responses and difficulties with file uploads, indicating a partial outage affecting the backend. The incident began at Jan 15, 2026 7:28 AM UTC when the status page moved from up to warn with the message 'Compaction is having issues' and a fix being implemented. The outage lasted approximately 9 hours and 35 minutes, with intermittent issues during that period.

Anthropic Claude experienced a global outage affecting the Claude API, Claude Code, and the Claude AI interface. Users reported prompt failures and overloaded errors across regions, with many messages describing timeouts and incomplete responses. The outage lasted about 3 hours and 20 minutes, from 10:48 AM UTC to 2:08 PM UTC. A substantial volume of user reports confirmed broad impact across multiple geographies.