Is OpenAI down today?

Check the status badge at the top of this page for the current status. Pulsetic polls the official status page every few minutes and cross-references with user reports to give you the most accurate picture.

How often does OpenAI go down?

Based on our 90-day data, we've detected 74 incidents, of which 8 were major or critical. The median resolution time is 1h 39m.

How are outages detected?

Pulsetic polls the official status page every few minutes from multiple global locations. Crowdsourced user reports provide an additional early-warning signal, often surfacing issues before the official monitors catch them.

Is it down for everyone or just me?

The outage map and region breakdown help answer this. If reports are concentrated in one area it may be regional; if hotspots appear globally it's a widespread outage. Submit your own report to help others understand the scope.

What does "never acknowledged" mean?

"Never acknowledged" means Pulsetic detected a real service disruption, confirmed by multiple user reports, but the official status page never posted an incident for it. This is one of the core reasons teams use third-party monitoring alongside official status pages.

ACTIVE INCIDENT DETECTED

OpenAI is
experiencing issues.

Pulsetic detects outages before the provider officially reports them. Monitoring since Apr 2026 · 74 incidents caught

Last checked: 1 min ago · 0 user reports (24h) · Trusted by 30,000+ teams

MINOR · INVESTIGATING

Codex, workspace analytics, conversation search, searching for custom GPTs, ChatGPT user invites, and Compliance Log Platform download endpoint not working in FedRAMP workspaces

Started 1 week ago · Jul 1, 2026 · 3:38 AM UTC

Core functionality has been restored but we are aware of known and ongoing issues in FedRAMP workspaces with Codex, workspace analytics, conversation search, searching for custom GPTs, ChatGPT user invites, and the Compliance Logs Platform download endpoint. We are working to resolve these issues and will share further updates we have more information available.

Having issues too?
User Reports: Last 24 Hours
0 reports

Bars represent 20-min intervals. Occasional spikes may reflect individual issues, not widespread outages.

Outage Map
No reports yet

Reports by region (last 24h)

No reports in the last 24h

View full map →
Recent User Reports
Live

Latest 50 reports from the last 48 hours

Top reported issues

Click an issue to upvote it and help us better indicate the service status.

Service outage 0

Degraded performance 0

Error messages 0

Timeouts / no response 0

Authentication issues 0

Network / connectivity issues 0

Feature not working 0

API errors 0

Billing or usage issues 0

Other 0

Have a different problem?
Issue breakdown
0 total votes

Distribution of reported issues over the last 24 hours.

Service outage 0%

Degraded performance 0%

Error messages 0%

Timeouts / no response 0%

Authentication issues 0%

Network / connectivity issues 0%

Feature not working 0%

API errors 0%

Billing or usage issues 0%

Other 0%

Last 90 days

90-Day Health Summary

Total Incidents

74

Since Apr 2026
Major / Critical

8

8 high-impact incidents
Minor Incidents

66

Avg. 0.7 per day
Median Resolution

1h 39m

P95: 32h 4m

OpenAI Response Time, ms

By Continent

Location List

From official status page

Recent Outage History

View all incidents →

Incident	Started	Duration	Impact	Status	Link
Users are experiencing elevated errors when selecting models	Jul 9, 2026 · 7:44 PM	58m	None	Resolved	Details →
Elevated errors with image generation in ChatGPT	Jul 7, 2026 · 11:29 AM	4h 43m	Minor	Resolved	Details →
Codex, workspace analytics, conversation search, searching for custom GPTs, ChatGPT user invites, and Compliance Log Platform download endpoint not working in FedRAMP workspaces	Jul 1, 2026 · 3:38 AM	Ongoing	Minor	Investigating	Details →
Issues with ChatGPT Windows desktop app	Jun 27, 2026 · 10:00 PM	5h 28m	Minor	Resolved	Details →
Codex Usage Limits Depleting Faster Than Expected	Jun 26, 2026 · 5:04 PM	3d 2m	Minor	Resolved	Details →
Small Number of Users Have Incorrectly Cancelled Subscription	Jun 25, 2026 · 4:02 PM	10h 15m	Minor	Resolved	Details →
Users may encounter issues using access token with Codex	Jun 25, 2026 · 5:51 AM	5h 12m	Minor	Resolved	Details →
gpt-4o-mini high error rate	Jun 24, 2026 · 5:59 AM	7m	Minor	Resolved	Details →
Users may experience elevated errors in ChatGPT uploading and downloading files	Jun 23, 2026 · 2:24 PM	2h 14m	Major	Resolved	Details →
chatgpt.com access issues	Jun 19, 2026 · 3:39 AM	1h 43m	Minor	Resolved	Details →

How do they compare?

Alternatives to OpenAI

Character AI

AI assistant

Online 16 incidents / 30d
Compare →
ChatGPT

AI assistant

Outage 20 incidents / 30d
Compare →
Claude

AI assistant

Online 41 incidents / 30d
Compare →

Teams monitoring OpenAI also watch

Other Monitored Services

Get alerted the moment OpenAI goes down.

No credit card · free plan · 2-minute setup

Start Free Trial

Common questions

Frequently Asked Questions

Is OpenAI down today?

Check the status badge at the top of this page for the current status. Pulsetic polls the official status page every few minutes and cross-references with user reports to give you the most accurate picture.
How often does OpenAI go down?

Based on our 90-day data, we've detected 74 incidents, of which 8 were major or critical. The median resolution time is 1h 39m.
How are outages detected?

Pulsetic polls the official status page every few minutes from multiple global locations. Crowdsourced user reports provide an additional early-warning signal, often surfacing issues before the official monitors catch them.
Is it down for everyone or just me?

The outage map and region breakdown help answer this. If reports are concentrated in one area it may be regional; if hotspots appear globally it's a widespread outage. Submit your own report to help others understand the scope.
What does "never acknowledged" mean?

"Never acknowledged" means Pulsetic detected a real service disruption, confirmed by multiple user reports, but the official status page never posted an incident for it. This is one of the core reasons teams use third-party monitoring alongside official status pages.