Anthropic Outage History

Anthropic experienced a 1-hour service incident where users encountered errors when uploading documents to Google Drive from Claude.ai, including the desktop app and Cowork features. The issue affected console.anthropic.com, api.anthropic.com, and claude.ai platforms. The Google Drive upload functionality was successfully restored and the incident was resolved.

Anthropic experienced a minor incident where users were unable to add credentials to vaults through the console.anthropic.com interface. The issue affected the credential management functionality for 19 minutes before being resolved with an implemented fix.

Claude Desktop's Cowork feature failed to start for some users on version v1.3036.0, preventing access to collaborative functionality. The issue was identified and resolved through a software update to version v1.3109.0, with users instructed to check for updates to restore service.

Anthropic's Opus 4.6 model experienced an elevated rate of errors affecting the console, API, and Claude.ai for 56 minutes. The issue was identified and a fix was implemented to resolve the outage. Service was restored after the problem was addressed.

Anthropic experienced elevated errors across Claude.ai, API, and Claude Code for 2.9 hours, causing login failures, API 500 errors, and preventing users from running scheduled tasks or downloading outputs. The Claude API recovered first, followed by stabilization of login services, with the incident fully resolved after implementing a fix for the identified issue.

Anthropic experienced degraded performance on usage and analytics admin API endpoints, causing intermittent 503 errors that affected claude.ai, api.anthropic.com, and console.anthropic.com. The root cause was identified and fixes were deployed through the pipeline over approximately 6 hours. The incident was fully resolved after monitoring confirmed the fix was effective.

Anthropic experienced a major outage affecting Claude.ai, the API, and console, with elevated login errors preventing users from accessing Claude.ai and Claude Code services. The incident lasted 57 minutes, from 15:31 to 16:19 UTC on April 13th, 2026. The issue was identified and resolved with a fix applied to restore normal login functionality across all affected services.

Anthropic experienced elevated errors affecting requests to Claude models across all products including claude.ai, API, and console for 22 minutes. The issue was identified as specifically affecting non-Opus models (all models except Claude Opus 4.6 continued working normally). The problem was fully resolved after the team implemented a fix for the affected model endpoints.

Anthropic experienced degraded performance on their Vaults feature affecting console.anthropic.com for 12 minutes. The issue was investigated and resolved without further details provided about the specific cause or impact scope.

Anthropic experienced elevated error rates affecting certain connectors for 43 minutes on April 9th, 2026. The issue impacted both claude.ai and api.anthropic.com services. The incident was resolved after investigation and remediation efforts.