Beyond Speed Limit, Another Incident: Claude has experienced 25 service disruptions in the past two weeks, with three major outages across product lines, making Opus 4.6 a highly affected area.

According to 1M AI News monitoring, the Anthropic official status page shows that out of Claude's 5 product lines, 3 are currently in a "Major Outage" state: claude.ai, Claude API, and Claude Code; platform.claude.com and Claude for Government are operating normally. As of the time of writing, there are two active unresolved incidents: Opus 4.6 elevated error rate (investigating since March 27, 14:59 Beijing time), and Cowork connection reset error (ongoing since March 25, with the official recommendation to restart the Claude Desktop app as a temporary workaround).

The status page records show that within the 15 days from March 13 to 27, Claude has experienced approximately 25 separate incident events, with no incidents reported on March 14, 15, 20, and 24. The event with the highest frequency is the Opus 4.6 related elevated error rate, occurring almost daily from March 17 to 27. On some dates, multiple concurrent incidents happened, with 4 incidents recorded on March 18. The types of incidents include model error rate spikes, claude.ai frontend errors, login system outages, MCP call failures, and response time delays. In terms of 90-day uptime, claude.ai is at 99.03%, Claude API at 99.11%, and Claude Code at 99.35%.

Combined with Anthropic's disclosed peak period quota adjustments on the same day, Claude's service is facing a dual pressure of increased demand and capacity.