OpenAI experiences another outage with ChatGPT, Codex, and APIs all going down simultaneously.
OpenAI’s APIs, ChatGPT, and Codex experienced higher error rates on Saturday, marking the fourth disruption within four days for the organization. According to OpenAI’s status page, there were elevated error rates across its APIs, ChatGPT, and Codex, with users encountering 503 errors labeled as “biscuit_baker_service_me_circuit_open,” which hindered requests from reaching OpenAI’s servers. The company shifted from investigating the issue to monitoring it within an hour, stating that it had implemented a solution and was observing for recovery.
This outage impacted all three key products of OpenAI at the same time. The API dashboard indicated 12 affected components, ChatGPT listed 15, and Codex noted four, with ChatGPT showing the lowest uptime of any OpenAI product in the last 90 days. This timing is unfortunate for a company that merged ChatGPT, Codex, and its API into a single platform under Greg Brockman in May with the goal of consolidating engineering resources.
The incident on Saturday was not an isolated event; OpenAI’s status history recorded elevated API error rates on Thursday, increased errors in Codex Review on the same day, and more elevated error rates in both ChatGPT and API image generation on Wednesday. This trend follows a pattern that has continued since autumn 2025, with the status page documenting around 166 incidents over approximately nine months, averaging about 18 per month. ChatGPT also faced a significant outage on April 20 and once more on July 23, just two days before Saturday’s incident.
The questions regarding reliability now weigh more heavily than they did a year ago. OpenAI’s products surpassed 10 million weekly users last week, and the launch of ChatGPT Work this month promises that AI agents can complete multi-step tasks across users’ applications and files, which relies on the platform’s stability. When a chatbot becomes unavailable, users lose their conversation; conversely, if an autonomous agent goes offline during a task, the repercussions can affect the entire workflow it was managing.
OpenAI is not the only organization facing these challenges. Anthropic's Claude experienced downtime in June, raising similar concerns about whether the infrastructure supporting advanced AI models can meet the growing demands placed upon it. However, the difference lies in scale and ambition: OpenAI currently serves 900 million weekly active users on a unified platform that it is preparing to take public later this year, and each outage undermines the reliability narrative that enterprise customers require before entrusting real tasks to an AI agent.
Other articles
OpenAI experiences another outage with ChatGPT, Codex, and APIs all going down simultaneously.
OpenAI's status page indicated increased error rates in its APIs, ChatGPT, and Codex on Saturday, marking the most recent in a series of outages throughout July.
