Elevated error rate on multiple models

Anthropic resolved an issue causing elevated error rates on multiple AI models, including Claude Opus 4.7, Sonnet 4.6, and Haiku 4.5, after a roughly four-hour incident on May 22. The company identified the problem at 04:40 UTC and reported success rates returning to normal on Opus 4.7 and Sonnet 4.6 by 05:53 UTC, with full resolution confirmed by 08:50 UTC.

May 22 , 08:50 UTC Resolved - This issue has been resolved. May 22 , 06:32 UTC Update - We are monitoring recovery on Claude Opus 4.7, and working to fully resolve remaining errors impacting Haiku 4.5. We will provide additional updates shortly. May 22 , 06:17 UTC Update - We are currently working to resolve some additional errors impacting Claude Opus 4.7 and Haiku 4.5, we will provide an additional update shortly. May 22 , 05:53 UTC Monitoring - We have seen success rates on Claude Opus 4.7 and Sonnet 4.6 return to normal. We are monitoring closely to ensure no further issues, and will provide an additional update shortly. May 22 , 05:25 UTC Update - We are continuing to work to fully resolve issues impacting Claude Opus 4.7, and Sonnet 4.6. At this time, success rates on Claude Opus 4.7 have stabilized. We will provide additional updates shortly. May 22 , 04:40 UTC Identified - We have identified an issue resulting in elevated error rates, primarily on Claude Opus 4.7 and Sonnet 4.6. We are working to resolve these issues, and will provide additional updates as soon as possible. May 22 , 04:16 UTC Investigating - We are currently investigating this issue.