ChatGPT & Codex Hit by Partial Outage: What You Need to Know

When your AI co-pilot suddenly goes dark, you notice. On April 20, 2026, users around the planet found ChatGPT and Codex stalled: blank screens, login failures, service errors. What’s behind the outage? How widespread is it? And when will things get back to normal? Here’s the latest.

What’s Going On — The Outage Unpacked

At approximately 8:05 PM UTC (that’s 20:05 on OpenAI’s timeline), OpenAI acknowledged a partial outage affecting both ChatGPT and Codex services. Around that time, users began seeing issues with gateway errors, login failures, blank pages, and general inability to access conversations. The disruption impacted 12 different ChatGPT components, plus at least one Codex component.

On Downdetector, user-reported issues surged from near-quiet levels to over 1,700 active reports by around 8:32 PM on April 20, up from a normal baseline in the low teens — a dramatic spike that clearly flagged a global problem, not just a regional hiccup.

How Dramatic Is the Fallout?

The outage wasn't just a momentary lag — services showed significant degradation. ChatGPT Business users reported delays, especially when adding new seats or making account-level changes. Many users — free or paid — saw complete failures when trying to log in or load content. The APIs & developer tools tied to Codex also flagged errors.

Based on current reports, the issue persisted for nearly 50 minutes by the time of the last official update, with impact zones including North America, parts of Europe, and Asia. While OpenAI began applying mitigation measures in hopes of speeding recovery, as of now some symptoms — especially login issues and blank responses — still remain in spots.

What Might Have Gone Wrong

OpenAI hasn’t released a full root cause yet. But based on the pattern — sudden onset, widespread component failure, and errors across core functionality — the likely culprits are configuration issues, infrastructure missteps, or newly deployed code with unintended side effects.

In previous outages earlier this year, similar symptoms emerged after updates or backend changes, especially when load-balancing or server routing was involved. Given that 12 distinct ChatGPT modules were affected, the chance of a centralized system error (rather than isolated bugs) seems high.

What You Can Do Now—and How to Check Status

If you’re experiencing issues with ChatGPT or Codex, here are your best moves:

• Visit OpenAI’s official Status Page to see live updates for ChatGPT, Codex, the API platform, and business tools.
• Check Downdetector or similar outage aggregators. A sharp vertical spike — from under 50 reports to over 1,000 in minutes — usually means a system-wide outage.
• Scan recent messages on social media platforms. When thousands of people report the same errors, it’s more than just your browser.

If you need to get work done while parts of the system are down, try switching devices (some users report mobile app access has better luck), or shift to offline tools if possible. Avoid making critical updates or account changes until full functionality is restored.

Where Things Stand Now

By the time of the most recent updates, OpenAI had applied mitigation efforts and was monitoring recovery. Some users reported that things were returning to normal — login started working again in certain regions, and conversations reloaded properly. But others still encountered blank pages or errors, particularly in regions where the spike was highest.

OpenAI has made no commitment yet on a full resolution time, saying instead that it’s investigating — as is customary during such incidents. For now, casual users might time out briefly, while power users and developers depending on Codex and APIs may continue to see issues.

Bottom line: this was not a total system crash, but it was serious — hitting many of the critical services people rely on.

Looking ahead, transparency is crucial. If outages continue to hit core features, OpenAI will likely face growing pressure to explain more clearly what kinds of changes could have caused the problem, and how to prevent them in future updates.

Conclusion

Outages like this one highlight that even world-class AI systems aren’t immune to technical failure. For now, millions around the globe are waiting for full restoration of ChatGPT and Codex. The best thing you can do: stay tuned to official channels, rely on status dashboards, and know that recovery efforts are well underway. Fortunately, these tools are resilient — we’ll be back online before long.