The ChatGPT Outage Wave of July 2026: What Happened, When, and What OpenAI Has Said

Reports on Downdetector surged rapidly from roughly 2,000 to 3,000 complaints, with about 80% tied to core app functionality . OpenAI confirmed the incident was resolved later the same day, and the status page marked it as normal .

Regional reach

Users reported disruptions across every major region:

• US and Europe
• India — especially heavy impact, with outage starting around 2:30 p.m. IST
• Japan, Australia
• UAE — OpenAI confirmed disruptions affecting the region
• Broader reports also came from Asia and the UK

Earlier outages in the same week

The July 25 event did not happen in isolation. It was the fourth service disruption in four days for OpenAI, as reported by TNW .

July 21 (Tuesday) — DALL-E / image generation outage

A major outage struck ChatGPT's built-in image generation (DALL-E) starting at 10:36 a.m. UTC. OpenAI reported "elevated errors" for image generation requests via both the ChatGPT UI and the API. The incident lasted approximately 14 hours before full recovery .

July 23 (Thursday) — degraded API performance

A separate incident caused "elevated error rates" in ChatGPT and API services. It started at 5:56 p.m. UTC and lasted about 1 hour 16 minutes. The severity was classified as "Degraded Performance" rather than a full outage . Some sources note that this coincided with a planned wave of model deprecations, which added confusion for developers .

Reported uptime percentages

OpenAI's own status data for the April–July 2026 window showed:

• ChatGPT: 99.86% uptime across its fifteen tracked components
• APIs: 99.99% uptime
• Codex: 99.98% uptime

Third-party monitor OutageDeck measured OpenAI at 99.68% uptime over the last 90 days as of July 25, 2026 .

Common root cause?

OpenAI has not identified a common root cause linking the July 21, July 23, and July 25 incidents. The company's status page updates described each disruption separately without indicating they stemmed from a single underlying issue. Several outlets noted the unusual frequency — four incidents in four days — but OpenAI's post-mortem communications have not publicly connected them .

What this means for users

For the 900 million weekly active users of ChatGPT and the countless applications relying on its API, this week of disruptions was a stark reminder of the fragility behind even the most polished AI services. OpenAI's 99.86% ChatGPT uptime figure, while high, allows roughly three hours of downtime per quarter — and the July incidents alone were on pace to consume that entire budget .