OpenAI confirmed a widespread service disruption on April 20, 2026, affecting its flagship artificial intelligence chatbot, ChatGPT, and its programming-focused model, Codex. The incident, which began at approximately 10:05 a.m. ET, left thousands of users unable to access key features, log into their accounts, or load existing conversations. According to the official OpenAI status page, the company initially categorized the event as degraded performance before upgrading the incident to a partial outage as the scale of the disruption became clearer.
The technical issues were first identified when users began reporting 5xx error codes and persistent loading screens. By 10:35 a.m. ET, OpenAI issued its first official statement, confirming that it was investigating reports of users being unable to load ChatGPT and Codex. As the investigation continued, the company noted that the disruption extended to various components of the ecosystem, including voice mode, image generation, and the mobile application interface. While some users could still access the web interface, many reported that their conversation history was unavailable, or that the system failed to respond to new prompts.
Data from the outage-tracking platform Downdetector indicated a sharp spike in reports shortly after 10:00 a.m. ET. Reports peaked at over 5,000 incidents within the first hour of the disruption. The impact was felt globally, with the highest concentration of reports originating from the United States, the United Kingdom, and India. In the United States alone, more than 1,600 individual complaints were logged by 11:00 a.m. ET. Users in India also reported significant downtime, particularly affecting developers relying on Codex for real-time coding assistance.
OpenAI’s status dashboard provided specific technical updates throughout the morning. At 11:13 a.m. ET, the company announced that it had applied a mitigation strategy and was monitoring the recovery process. A specific advisory was issued for ChatGPT Business subscribers, noting that those attempting to upgrade their plans or add new seats might experience delays or errors for up to an hour following the initial mitigation. Despite the issues with ChatGPT and Codex, OpenAI reported that its core API services maintained a 99.99% uptime during the window, and its video-generation model, Sora, remained largely unaffected with 99.95% uptime.
As of the latest update, OpenAI has not disclosed the root cause of the infrastructure failure. The company continues to monitor system stability as services gradually return to normal for the majority of the user base. This event marks one of the most significant disruptions for the platform in the first half of 2026, highlighting the ongoing challenges of maintaining high-availability AI infrastructure for a global user base.