Developer frustrated by service outage, laptop screen shows "Service Unavailable."
OpenAI, a leading artificial intelligence research and deployment company, is currently experiencing a partial outage that has affected its widely used chatbot, ChatGPT, and its agentic coding platform, Codex. The disruption has led to access issues for thousands of users worldwide, as reported by outage tracking website Downdetector.
The company has acknowledged the issue and stated that it is under investigation. The outage is characterized as partial, with a range of features impacted, including conversational abilities, user login, voice mode, and image generation. This means that even users who can successfully log in may encounter errors when attempting to utilize specific functionalities.
Reports of the service disruption began to surface around 7:35 PM IST. Downdetector data indicates a notable spike in user complaints, with a significant number originating from India. At the peak of the outage, over 800 users reportedly logged issues related to ChatGPT, Codex, the mobile application, and login access, affecting both web and mobile platforms.
The specific cause of the current disruption has not yet been disclosed by OpenAI. The company also noted that it has implemented mitigation measures for a separate issue concerning ChatGPT Business users who were attempting to upgrade plans or add seats, and is actively monitoring the recovery process for that issue.
This incident follows a period of heightened reliance on AI services and comes months after a major global outage, attributed to Cloudflare, which impacted numerous platforms including ChatGPT, X (formerly Twitter), Spotify, and PayPal. That event underscored the critical dependence of digital services on a few key infrastructure providers.
For OpenAI, this marks the second significant outage this year. In March, ChatGPT experienced a widespread outage affecting thousands of users, with issues peaking around 2 AM IST. At that time, OpenAI cited elevated error rates impacting conversations and related services before operations were restored. The root cause was not disclosed.
About the Author
