OpenAI services were hit by a widespread outage on August 31, 2026, and down detector showed the pressure building fast. By 09:32 PM IST, the complaint count had climbed to 678 problem reports, with users describing unresponsive prompts, failed connections, and sudden network timeouts.
The disruption started showing up around 05:28 PM IST, when roughly 31 reports appeared, then spread through the evening as more users said OpenAI tools would not respond. That pattern matters for anyone relying on ChatGPT Work for text generation, workflows, or automated tasks, because OpenAI said users on multiple subscription plans may be unable to start or continue work.
ChatGPT Work errors and latency
OpenAI said it was experiencing issues across its infrastructure and that ChatGPT Work was seeing elevated errors and latency. The company also said its team was working on resolving the issue, which is the only response posted in the public tracker during the outage window.
For affected users, the practical problem was not just slow service. The reports pointed to failed connections and prompts that would not load or complete, which can interrupt live drafting, repeated task runs, and any workflow that depends on successive requests rather than a single reply.
Downdetector’s evening spike
Downdetector’s complaint curve kept rising throughout the evening, turning the outage into a visible service event rather than isolated user noise. Similar crowd-sourced spikes have appeared in other service disruptions, including a recent Reddit outage report that drew nearly 29,000 reports in 15 minutes, but OpenAI’s pattern here was tied to a narrower set of infrastructure and latency complaints.
That matters because the problem was not limited to one product tier or one type of task. OpenAI said multiple subscription plans could be affected, and the disruption reached the tools people use to generate text and manage automated work, not just casual chat sessions.
OpenAI’s tracker versus X
OpenAI acknowledged the disruption on its official incident tracker, but it had not posted formal updates or troubleshooting advice on its official X account. That gap leaves users with the tracker as the only public place where the company addressed the outage while the reports were still climbing.
The unanswered piece is what triggered the server breakdown. OpenAI has said its team is working on the issue, and the next meaningful signal for users will be when the tracker changes from active errors to a restored service state.
