Anthropic Outage History

Anthropic experienced elevated connection reset errors in Claude Cowork that lasted 48.5 hours. The issue affected Cowork sessions and could be temporarily resolved by restarting the Claude Desktop application. The incident was fully resolved on March 27th, 2026.

Anthropic's claude.ai experienced elevated errors for 2 hours, preventing users from accessing the service with messages like "this isn't working right now" and "this is not available right now," while also making saved chats inaccessible. The issue affected users globally, including those in the UK, and was resolved after engineers identified and implemented a fix.

Anthropic experienced elevated errors on Claude Opus 4.6 that affected claude.ai, console.anthropic.com, and api.anthropic.com for 3.6 hours. The company investigated the issue, implemented a fix, and monitored the results before fully resolving the incident.

Between 9:10 PT and 9:26 PT on March 23rd, Claude.ai experienced elevated error rates for approximately 16 minutes, affecting users' ability to interact with the web interface. The Claude API remained fully operational throughout the incident. The issue was resolved and service was restored to normal functionality.

Anthropic experienced a major service incident where users were unable to log in to Claude.ai and the Claude Console (platform.claude.com). The authentication issue affected both platforms for 21 minutes before a mitigation was successfully applied. The incident was fully resolved with logins and authentication returning to normal operation.

Claude.ai experienced delayed response completion where messages would hang for approximately 5 seconds after text finished streaming, starting around 20:00 UTC on March 20. The issue only affected the Claude.ai web interface while the Anthropic API continued to function normally. The incident was resolved after 2 hours once the issue was identified and a fix was implemented.

Anthropic experienced elevated error rates affecting Claude Opus and Sonnet 4.6 models across claude.ai, console.anthropic.com, and api.anthropic.com for 1.6 hours. The team identified the underlying issue and restored Claude Opus 4.6 to normal success rates first, followed by full resolution of errors for Claude Sonnet 4.6. The incident was fully resolved with all services returning to normal operation.

Anthropic experienced elevated errors on Claude Opus 4.6 affecting both claude.ai and api.anthropic.com for 20 minutes on March 19, 2026. The incident began at 8:53 PT and was resolved by 9:14 PT, with error rates returning to normal baseline levels.

Anthropic experienced elevated errors on Claude Sonnet 4.6 starting at 8:23 PT / 15:23 UTC, affecting both claude.ai and api.anthropic.com. The incident lasted 11 minutes and was resolved by 8:42 PT / 15:42 UTC when error rates returned to baseline levels.

Anthropic experienced elevated authentication errors across Claude.ai, console.anthropic.com, and api.anthropic.com from 23:59–00:30 UTC on March 19, 2026. Users were unable to authenticate properly when trying to access various Claude surfaces including Claude.ai and Claude Code. The authentication issue was resolved after 1.2 hours, with Anthropic monitoring the systems to ensure stability.