Anthropic Outage History

Anthropic Claude experienced elevated errors affecting Claude 4 Opus requests across the API, Claude.ai, and the Anthropic Console. The issue began at May 26, 2025 6:33 PM UTC and prompted ongoing investigations across three components: api.anthropic.com, claude.ai, and console.anthropic.com. Users may have encountered failed requests or degraded responses during the incident while updates were published.

The provider issued four status updates as the incident progressed, and the service returned to normal by May 27, 2025 12:38 AM UTC. The issue involved multiple Claude services globally and affected Claude 4 Opus functionality and code execution tools, before full restoration was confirmed. The outage lasted roughly 6 hours and 5 minutes.

From May 26, 2025 2:48 PM UTC to May 26, 2025 3:18 PM UTC, Claude Opus requests experienced elevated errors, affecting Claude 4 Opus usage across console.anthropic.com, api.anthropic.com, and claude.ai. Users attempting to use Claude during this window may have encountered failed requests or degraded performance.

The Claude team posted two status updates and indicated a fix had been implemented with monitoring underway. The outage lasted about 30 minutes, and services returned to normal by May 26, 2025 3:18 PM UTC.

The Claude Sonnet 4 service experienced elevated errors beginning at 5:30 AM UTC, impacting claude.ai, console.anthropic.com, and api.anthropic.com. Users attempting to use Claude Sonnet 4 may have encountered errors or degraded responses. The outage lasted about 5 hours and was resolved by May 23, 2025 10:33 AM UTC.

The provider issued two official status updates during the incident, with the status page indicating investigation starting at May 23, 2025 8:28 AM UTC and the services returning to normal by May 23, 2025 10:33 AM UTC.

Anthropic Claude experienced a brief, feature-specific outage affecting two areas: image usage in Claude Sonnet 4 conversations and tool usage in Claude Code on Opus 4. Users reported that images could not be displayed or attached in Sonnet 4 chats, and that Claude Code on Opus 4 produced errors when using tool features. The incident began at 7:43 PM UTC and impacted two components: claude.ai and Claude Code. The outage lasted about 2 hours and 15 minutes.

Anthropic experienced a brief partial outage affecting api.anthropic.com, with elevated errors on some models. Users attempting model requests may have seen failures or degraded performance from Claude AI and the Console, while the API endpoint was down temporarily. The incident began at May 21, 2025 11:33 PM UTC and was resolved by May 21, 2025 11:43 PM UTC.

The provider issued three status updates during the incident, describing ongoing investigation, implementation of a fix, and recovery of all components to normal operation.

Anthropic experienced a partial outage starting at May 21, 2025 6:33 PM UTC when claude.ai was listed as degraded due to MCP tool calling issues for Advanced Research queries. The incident then affected console.anthropic.com and api.anthropic.com, which showed elevated errors by May 21, 2025 8:23 PM UTC. Some users may have encountered failures or slower responses when using Claude AI models during this period.

Over the course of the incident, the provider issued four official status updates. By May 21, 2025 8:47 PM UTC the situation was described as Partial Degraded Service, with claude.ai beginning to recover and the other components trending toward normal. The outage lasted roughly 4 hours and 55 minutes, and service for claude.ai was restored by May 21, 2025 11:28 PM UTC. Affected components included claude.ai, console.anthropic.com, and api.anthropic.com.

Anthropic Claude API and Console experienced an incident with elevated errors on requests to certain models. The issue began at May 21, 2025 5:08 PM UTC and affected two components: api.anthropic.com and console.anthropic.com, limiting access to several Claude models. Some users reported issues earlier around 4:45 PM UTC. The impact included degraded ability to perform model calls and use the web console.

Anthropic issued two status updates during the incident. The initial update indicated that teams were investigating the problem, and the service was restored by May 21, 2025 6:26 PM UTC, when status returned to up. The outage lasted approximately 1 hour and 18 minutes and represented a focused disruption on two Claude endpoints.

Anthropic Claude users experienced elevated errors across multiple models, affecting three components: console.anthropic.com, api.anthropic.com, and claude.ai. The incident began at 10:05 AM UTC with degraded performance and a status update to degraded at 10:31 AM UTC. The issue impacted core interactions with the Claude platform during the outage.

The outage lasted about 1 hour and 8 minutes and ended when the status returned to normal at 11:13 AM UTC. The provider issued 2 status updates during the incident describing the elevated errors and subsequent recovery, with the three affected components returning to normal after the incident.

The Anthropic API (api.anthropic.com) experienced a brief degradation with elevated errors for web search requests routed through the web search server tool. The outage began at 1:33 AM UTC and was resolved by 1:43 AM UTC, lasting about 10 minutes. The provider issued two status updates during the incident, noting that the issue was identified and a fix was in progress. The impact was limited to the API endpoint and did not indicate a broader regional outage.

The Anthropic Console briefly experienced elevated errors when users attempted to purchase API credits, impacting the credits purchase flow. The incident began at May 17, 2025 3:58 PM UTC and was resolved by May 17, 2025 7:48 PM UTC, lasting roughly 3 hours and 50 minutes. Anthropic's status page stated that the issue had been identified and they were working to resolve it as the console gradually returned to normal.