四大主要AI模型服务商同时遭遇罕见中断

OpenAI、Anthropic、xAI和Google四家主要AI模型服务商周四上午遭遇罕见的同时服务中断。1Anthropic的Claude系列模型最早在美东时间9:23报告错误，约15分钟后识别出问题根源，并于12:16修复。1受影响的Claude型号包括Claude Mythos 5.1、Claude Fable 5.1、Claude Opus 5和Claude Sonnet 5。1OpenAI的ChatGPT和Codex服务从10:43开始出现性能下降，该公司在约半小时后部署了缓解措施，并于12:55完全恢复。1

根据Downdetector的报告统计，OpenAI收到超过12000份问题报告，Claude约1200份，Grok约1000份。2Codex用户遭遇404 Not Found错误。2Anthropic表示正在调查Claude错误并进行修复，2而SpaceXAI确认Grok出现故障并正在调查服务中断。2各服务商的问题在9月4日1:13已恢复正常。2

Four leading artificial intelligence service providers experienced simultaneous outages on Thursday morning, disrupting access to some of the world's most widely used AI platforms.1 Anthropic's Claude model suite began reporting errors at 9:23 a.m. Eastern Time, with the company identifying the underlying cause approximately 15 minutes later and resolving the issue by 12:16 p.m.1 The affected Claude models included Claude Mythos 5.1, Claude Fable 5.1, Claude Opus 5, and Claude Sonnet 5.1 OpenAI's ChatGPT and Codex experienced performance degradation starting at 10:43 a.m., with the company deploying mitigation measures roughly half an hour later and fully restoring service by 12:55 p.m.1

The scale of the incident was substantial, with incident tracking services recording over 12,000 problem reports for OpenAI, approximately 1,200 for Claude, and around 1,000 for Grok.2 Users of Codex encountered 404 Not Found errors during the outage.2 Both Anthropic and OpenAI confirmed they were investigating their respective service interruptions, with Anthropic stating it was working to repair Claude errors while OpenAI acknowledged issues affecting ChatGPT and Codex.2 By September 4 at 1:13 a.m., all affected services had returned to normal operation.2