What Happened
On the day of the incident OpenAI reported two concurrent problems: “Elevated error rates for finetuning jobs” and “Elevated errors for ChatGPT and Platform users.” The issues caused delayed responses, occasional non‑responses, and slower performance across both free and Plus accounts.
Impact on Users
Users experienced a range of symptoms, including:
- ChatGPT not responding at all on free accounts.
- Slower reply times on Plus accounts.
- Higher error messages during finetuning operations.
Down Detector recorded more than 12,000 reports in the United States before the count fell to roughly 3,700 as the situation improved.
OpenAI's Response
OpenAI confirmed the first issue about 20 minutes after it began and the second roughly 44 minutes later. The company posted an update on its status page stating:
“We have identified the issue, applied the necessary mitigations and are monitoring the recovery.”
Technical teams applied mitigations to the error‑prone services and began monitoring the platform for stability.
Current Status
By early afternoon (ET), the majority of reports indicated that ChatGPT was back to normal operation in the United States. Users in the United Kingdom reported intermittent success, with many finding that accessing the service via a US‑based VPN restored full functionality.
Workarounds and Recommendations
- Log out of ChatGPT and log back in to refresh the session.
- Use a VPN set to a region where the service is confirmed stable (e.g., United States).
- Monitor OpenAI’s official status page for real‑time updates.
Conclusion
The incident highlights the importance of rapid communication and transparent status reporting during service disruptions. OpenAI’s swift acknowledgment and mitigation efforts helped reduce the outage’s impact, and the platform appears to have returned to normal operation for most users.