on may 27, openai officially confirmed that its core services experienced a global response delay: both the chatgpt web interface and the api showed significant performance degradation, with users widely reporting longer response times and slower request processing. according to real-world testing by users in multiple regions, since the early hours of the 27th (beijing time), query wait times have noticeably increased, and third-party integrated applications have been similarly affected.
at 10:47 pm beijing time, openai upgraded the incident to “high latency” on its status page and activated its emergency response mechanism. after approximately five hours of technical troubleshooting, the main service regained stability at 4:06 am the following day. however, two issues remain to be optimized: the codex model’s context compression efficiency has not yet reached baseline levels, and the android enterprise edition of chatgpt occasionally encounters anomalies when switching between multiple workspaces; fixes are still underway.
this disruption occurred at a critical juncture as openai accelerates its commercialization efforts—just as it is restructuring its advertising system and rolling out tiered subscription plans. response stability directly impacts developer ecosystem trust and end-user experience. although the outage was relatively brief, it has prompted the industry to closely examine the underlying architecture’s fault tolerance and operational response efficiency. currently, all services have returned to normal, with all core features functioning without interruption. the platform continues to strengthen its real-time monitoring and automated alert systems to enhance overall system resilience.
