Space for ads
Checked every 5 minutes · How we measure this
As of 23:50 EDT on 21 Sep 2026, chatgpt.com responded to our check in 158 ms, with 100.00% availability over the last 24 hours.
Checked every 5 minutes · last checked 7 minutes ago
The line is response time measured from our probe to chatgpt.com, which rises when the service slows down. Shaded red marks intervals where checks actually failed. It excludes your own connection and browser rendering.
Reported by ChatGPT on its own status page. Verify on their status page.
0 problem reports
SSO sign-in and SCIM provisioning issues
We have applied the mitigation and are monitoring the recovery.
Confirmed on the vendor status page
Overbilling for OpenAI-hosted containers in the Agent API
We have applied the mitigation, and any new sessions would not run into this issue. We are monitoring the recovery.
Confirmed on the vendor status page
Delayed support responses
We have applied the mitigation and are monitoring the recovery. Some support responses by email and chat may still be delayed while we work through affected cases.
Confirmed on the vendor status page
We are seeing elevated error rates across API models
We have applied the mitigation and are monitoring the recovery.
Confirmed on the vendor status page
Elevated errors affecting ChatGPT Work mode
We are still implementing the mitigation.
Confirmed on the vendor status page
Elevated errors in ChatGPT Work
We have applied the mitigation and are monitoring the recovery.
Confirmed on the vendor status page
ChatGPT degraded performance
DownBits detected degraded performance on ChatGPT between 16 Sep 2026, 05:30 – 06:10 IST, lasting ongoing. Detected across 8 consecutive automated checks of chatgpt.com.
Detected by DownBits probe
Elevated errors with gpt-image-2.5-flare
We have applied the mitigation and are monitoring the recovery.
Confirmed on the vendor status page
Elevated errors from GPT-5.6 and GPT-5.6 Instant on paid plans
We have applied the mitigation and are monitoring the recovery.
Confirmed on the vendor status page
Degraded Performance affecting Agents API
We have applied mitigations and are seeing recovery, but the service has not fully recovered. We are continuing to monitor recovery.
Confirmed on the vendor status page
Elevated errors affecting Work Mode in ChatGPT
We have applied additional mitigations for Work Mode. We are monitoring to confirm that workspace tools and files are functioning normally for ChatGPT Plus users.
Confirmed on the vendor status page
Elevated error rates for Codex and ChatGPT Work
We have applied the mitigation and are monitoring the recovery.
Confirmed on the vendor status page
Elevated errors for ChatGPT users in Europe
We have applied the mitigation and are monitoring the recovery.
Confirmed on the vendor status page
ChatGPT degraded performance
DownBits detected degraded performance on ChatGPT between 13 Sep 2026, 15:05 – 15:30 IST, lasting ongoing. Detected across 5 consecutive automated checks of chatgpt.com.
Detected by DownBits probe
Codex GitHub Review and Pull Request Failures
Github have applied the mitigation and are monitoring the recovery.
Confirmed on the vendor status page
ChatGPT is responding normally. Our last automated check of chatgpt.com completed 7 minutes ago and succeeded in 158 ms. Over the past 24 hours the service answered 100.00% of our checks.
Measured from our automated checks, ChatGPT has been available 100.00% over 24 hours, 99.90% over 7 days, 99.90% over 30 days. We recorded 26 incidents in the last 30 days.
ChatGPT has averaged 170 ms across our successful checks in the past 24 hours, which is fast — comfortably within what a healthy service should return. Response time is measured from our probe to chatgpt.com and does not include rendering time in your browser.
The most recent incident DownBits recorded for ChatGPT began on 16 September 2026 at 05:30 IST and lasted 40 min. It was logged as degraded performance. Detected by DownBits probe.
Our probe is reaching chatgpt.com normally right now, so the problem is more likely to be local to you than a full outage. Use the report button on this page to add your own signal — when several people report the same problem within a short window, it appears on the chart above.
A working marketing site does not mean a working API. Call the API host directly — curl -s -o /dev/null -w "%{http_code}" https://chatgpt.com/ — and check the vendor status page linked above for component-level detail, since one component can fail while the rest stay green.
HTTP 429 and 529 mean you are being throttled, not that the service is down. Inspect the response headers for a retry-after value and back off exponentially rather than retrying immediately.
A 401 or 403 is an authentication problem on your side. Rotated keys, an expired billing method, or a revoked token all present as sudden total failure.
Open any other site. If nothing loads, the issue is your network rather than ChatGPT. On Windows run ping chatgpt.com and nslookup chatgpt.com in Command Prompt; on macOS or Linux use Terminal. A DNS failure and a server failure look identical in a browser but are very different problems.
Load chatgpt.com in a private or incognito window. If it works there, clear cookies and site data for chatgpt.com in your normal browser rather than clearing everything.
Switch from Wi-Fi to mobile data, or vice versa. Regional ISP routing problems are common and often affect one carrier while others are unaffected.
We link the vendor status page so you can check our reading against theirs.
No comments received
