Good news, vibe coders: OpenAI says Codex is back to normal after experiencing 'elevated errors'

Codex vibe coders can get back to work after some hiccups earlier this morning.

OpenAI acknowledged that Codex was experiencing “elevated errors” as users complained about issues with the AI coding tool. “Oy,” OpenAI’s Codex lead Thibault Sottiaux wrote on X. “We are aware that some Codex users are experiencing high error rates with ‘model at capacity’ and are working to bring things back to being stable.” The initial error on OpenAI’s dashboard was first listed at 6:32 a.m.

Eastern Time on Tuesday, OpenAI’s status indicator had changed to green with the message “All impacted services have now fully recovered.” Across social media, Codex users shared screenshots of their Codex dashboards not working.

Screenshots showed that GPT 5.5, its current frontier coding model, was “at capacity” and that users needed to use an older model.

I’m facing this issue in codex again and again, anyone else?

Are we getting GPT 5.6?@thsottiaux @reach_vb @jxnlco pic.twitter.com/iJ2ECZDncp “It seems Codex is busy today.

I got several ‘Model at capacity’ errors 🤔,” Tomás Ruiz-López, a principal software engineer at GoodNotes wrote on X.

While Anthropic’s Claude Code dominates the AI coding market, OpenAI has positioned Codex to chip away at that lead.

Both rival AI companies have publicly confirmed they have taken the first steps toward their highly anticipated IPOs.

A spokesperson for OpenAI did not immediately respond to a request for comment from Business Insider.