Incident history

OpenAI — past outages

Verified against our probes + official status. Last 12 months.

← Live status

Uptime 12mo

99.91%

Incidents

Major

Total downtime

7h 52m

All severities 🔴 Major 🟠 Minor 🔧 Maintenance

Date	Incident	Severity	Duration	Scope
Jun 15, 2026 ongoing	Elevated error rates — Chat Completions API	minor	—	EU, SEA
May 28, 2026	Major outage — API & ChatGPT unavailable	major	2h 15m	Global
May 03, 2026	Scheduled maintenance — model deployment	maint	40m	API
Apr 19, 2026	Increased latency on fine-tuning endpoints	minor	1h 05m	US East
Mar 22, 2026	Authentication failures (API keys)	major	1h 30m	Global
Feb 14, 2026	DALL·E image generation errors	minor	52m	Global

Notable incident

May 28, 2026 — Major API & ChatGPT outage

Resolvedmajor

17:05 UTCDetected

Probes from all 7 regions began returning 5xx on api.openai.com.

17:22 UTCProvider acknowledged

Official status page moved to "investigating".

18:40 UTCPartial recovery

ChatGPT web restored; API still degraded in EU.

19:20 UTCResolved

All endpoints operational across regions.

✨ AI-generated incident summary

On May 28, 2026, OpenAI experienced a global outage lasting 2h 15m. Both the API and ChatGPT returned 5xx errors starting 17:05 UTC, first detected by dudown probes 17 minutes before the official acknowledgement. The root cause was attributed to an upstream infrastructure provider. ChatGPT web recovered first (18:40 UTC); the API fully recovered by 19:20 UTC. Apps using completions with no fallback would have seen total failure during the window.

Impact at a glance

Endpoints affectedAPI + ChatGPT

Peak user reports14,300 / hr

Lead time vs official17 min earlier

⚠️ Mock / for reference only. AI-generated summaries are illustrative and may be imperfect; verify against the provider's official post-mortem. Status and dates shown are fictional sample data.