Anthropic’s Claude platform suffered a major outage on April 15, 2026, knocking out access to Claude.ai, the developer API, and Claude Code for nearly three hours. The company confirmed it was investigating elevated errors at 7:53 PM Pakistan Standard Time (10:53 AM ET), and the incident was not fully resolved until 10:42 PM Pakistan Standard Time (1:42 PM ET).
Recovery was uneven throughout the outage. At 8:20 PM Pakistan Standard Time (15:20 UTC), Anthropic confirmed the API had fully recovered, but Claude.ai remained down and new logins were still broken. An update at 8:40 PM Pakistan Standard Time stated bluntly that “Claude.ai and Platform are down,” with Claude Code logins also failing. By 9:29 PM Pakistan Standard Time, login success rates had stabilized, and the incident was marked resolved at 10:42 PM Pakistan Standard Time.
Outage tracking site Downdetector recorded more than 7,000 user reports within the first hour, surpassing 15,000 within 25 minutes of that and eventually exceeding 20,000. Among those affected, 46% cited problems with Claude Code, 24% reported issues with Claude Chat, and 18% pointed to the mobile app. Google search interest in “Is Claude down” surged 500% during the outage window. One widely shared social media reaction read: “Claude Code is down right now and suddenly half the software industry has to remember how to think.”
Pro subscribers found the outage particularly frustrating, as the system could not generate or process verification codes, locking them out entirely. Claude Code users who were already logged in could continue working, but new sessions were blocked. Free users were unable to access the service across desktop and mobile.
The April 15 incident fits a pattern of recurring disruptions. Outages hit on April 3, 6, 7, 8, and 13, before this latest one. The frequency points to growing infrastructure strain as demand for Claude continues climbing.
The outage also arrives amid wider complaints about a perceived decline in Claude’s response quality, with developers reporting the model increasingly fails to follow instructions and takes shortcuts on complex tasks. Anthropic declined to comment on those claims. The company was valued at $380 billion as of February 2026 and is reportedly preparing for an IPO.
