Today, Anthropic faced a significant disruption with its Claude AI models, affecting services such as Claude Code. Users attempting to access Claude Code were met with a 500 error, while Anthropic reported increased error rates across all Claude model APIs.

The company quickly pinpointed the issue and implemented a solution within about 20 minutes. However, during this time, developers who rely on the service were left waiting for operations to resume. In addition to this incident, Claude Opus 4.5 experienced some errors yesterday, and earlier in the week, Anthropic addressed problems related to the purchase of AI credits.

Such a substantial outage is rare for Claude Code, particularly given its extensive use among developers, including Microsoft’s AI teams.

Update, February 3rd: Anthropic has successfully resolved the Claude API outage.