OpenAI's Codex Recovers After Experiencing Elevated Errors
OpenAI's Codex tool experienced "elevated errors" and "at capacity" messages on Tuesday morning, leading to user complaints. The issues began around 6:32 a.m. Eastern Time, with users unable to access the AI coding tool. OpenAI acknowledged the problem and worked to resolve it, confirming full recovery of all impacted services by approximately 9:50 a.m. Eastern Time.
OpenAI's AI coding tool, Codex, encountered "elevated errors" and displayed "at capacity" messages on Tuesday morning, according to user reports. The problem, which started around 6:32 a.m. Eastern Time, prevented some users from utilizing the service.
OpenAI's Codex lead, Thibault Sottiaux, acknowledged the high error rates and the "model at capacity" issue, stating that the company was working to stabilize the service. Users reported seeing messages that its current frontier coding model, GPT 5.5, was at capacity, suggesting the use of an older model.
The service fully recovered by approximately 9:50 a.m. Eastern Time. OpenAI's status indicator turned green at 9:48 a.m. Eastern Time, accompanied by a message confirming that "All impacted services have now fully recovered.
The AI coding market is competitive, with Anthropic's Claude Code a significant player. OpenAI has positioned Codex to compete in this sector. Both OpenAI and Anthropic have publicly confirmed initial steps toward their highly anticipated IPOs.
(Source: Business Insider)
