Anthropic restored Claude on Sunday, August 16, 2026, after an outage disrupted authentication and degraded several of its AI services, including claude.ai, the Claude Console, Claude API, Claude Code and Claude Cowork. The company began investigating at 21:58 UTC and marked the incident resolved at 22:34 UTC, putting the official incident window at about 36 minutes.
The disruption mattered beyond people using Claude as a chatbot. Anthropic’s API and coding tools are used in software and automated workflows, so problems affecting several services at once can interrupt development work as well as ordinary conversations.
Claude outage moved quickly from login trouble to broader service disruption
Anthropic’s incident record shows the problem first appeared as an authentication issue affecting claude.ai, Claude Code and Claude Cowork. At 22:02 UTC, the company expanded the notice to degraded performance across claude.ai, platform.claude.com, the Claude API, Claude Code and Claude Cowork.
A fix was deployed at 22:22 UTC, and Anthropic said it was monitoring recovery. Twelve minutes later, at 22:34 UTC, the company marked the incident resolved. The full timeline remains available on Anthropic’s official Claude incident page.
Anthropic did not publish a detailed root cause for the August 16 disruption in that incident record. The company described the authentication problem and wider degradation, but it did not attribute the outage to a cyberattack, data breach or loss of user information.
Why the outage affected more than chatbot conversations
Claude Code is designed for software-development work, while the Claude API allows applications to send requests directly to Anthropic’s models. When those services become unavailable, developers can see failed requests, interrupted coding sessions or delays in automated tasks even if their own systems remain online.
This was not Claude’s only reliability issue in 2026. A June Claude outage affected the chatbot, API and Claude Code, while an earlier Claude API outage produced 500 errors and thousands of user reports in April.
Those earlier incidents should not be treated as evidence of one recurring fault. Similar symptoms can come from different parts of a cloud service, and Anthropic has not publicly linked the August 16 disruption to the June or April outages.
Separate Claude model issues followed later in the week
Anthropic’s status history shows that separate incidents followed after the August 16 outage. On August 17, the company investigated elevated errors involving Claude Opus 5 and Claude Sonnet 5 before resolving the issue later that day.
On August 18, Anthropic reported degraded performance across multiple models, with the remaining impact on Claude Opus 5 resolved after errors had affected requests between 16:11 and 18:23 UTC. Another incident on August 19 involved Claude Opus 5 and Claude Haiku 4.5 and was resolved at 11:02 UTC.
These later events were listed as separate incidents, so they should not be described as a continuation of the August 16 authentication outage. As of August 20, Anthropic’s status page showed claude.ai, Claude Console, the Claude API, Claude Code, Claude Cowork and Claude for Government as operational.
What users and developers can take from the incident
For users who suddenly cannot sign in or send prompts, checking the provider’s status page first can show whether the problem is platform-wide before they spend time changing passwords, browsers or devices.
For developers, short outages are a reminder to handle temporary API failures carefully. Reasonable retry limits, clear error messages and fallback plans for critical workflows can reduce the impact when an external AI service is briefly unavailable.
The August 16 outage was resolved in less than an hour, but the separate model incidents that followed on August 17, 18 and 19 show why service status should be judged incident by incident. By August 20, Anthropic was reporting all listed Claude services as operational.















