The official status page of OpenAI shows that ChatGPT is currently experiencing service failures, with increased error rates, some users unable to log in normally, and intermittent errors. The OpenAI team is investigating and handling the issue. As of the time of this report, the failure has lasted more than 20 minutes.

GPT-5.6 three-tiered product line has just been launched

This failure occurred shortly after the official launch of the GPT-5.6 series model. OpenAI recently released three differentiated models using a new celestial naming hierarchical system: the flagship GPT-5.6 Sol (Sun), which focuses on complex reasoning, code development, scientific research, and cybersecurity scenarios; the balanced GPT-5.6 Terra (Earth), suitable for daily mainstream office tasks, with performance comparable to GPT-5.5 but at half the cost; and the lightweight GPT-5.6 Luna (Moon), focusing on response speed and low cost, suitable for high-frequency batch calling scenarios.

Industry opinions suggest that the three-tiered layout of GPT-5.6 is a strong counterattack by OpenAI against its competitor Anthropic, covering different customer segments through differentiated pricing. Combined with the ChatGPT Work intelligent agent product, OpenAI aims to gradually shift the core competition focus in the AI field to its more advantageous consumer market.

ChatGPT is not the first time experiencing large-scale failure

This is not the first time ChatGPT has experienced a large-scale service anomaly. In December 2023, ChatGPT also experienced a similar major failure, with nearly 3,000 service interruption reports recorded by the monitoring platform Downdetector at that time. As one of the AI applications with the largest user base globally, each failure affects a large number of users who rely on it. Service stability remains a challenge that OpenAI needs to continuously address while rapidly iterating.