AnswersPublished22 sources
ChatGPT Outage Reports on Sept. 25: What Downdetector Saw and OpenAI Confirmed
On Sept. 25, Downdetector logged more than 2,700 ChatGPT problem reports by 3:55 p.m. September’s status history also lists a Sept.
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer
Key takeaways
- On Sept. 25, Downdetector logged more than 2,700 ChatGPT problem reports by 3:55 p.m.
- September’s status history also lists a Sept. 3 ChatGPT and Codex disruption and later issues involving paid plan models, API models and ChatGPT Work.
- The supplied OpenAI status snapshot lists 99.96% API uptime, not 99.93%; aggregate uptime does not mean every feature or user was unaffected.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: What happened during the Friday, September 25, 2026, disruption affecting ChatGPT users, how did Downdetector reports compare with OpenAI’s. Article summary: On Friday, September 25, users reported trouble accessing ChatGPT, including “Unexpected status 401: Incorrect API key” errors. Downdetector had more than 2,700 reports by 3:55 p.m. Pacific, while OpenAI’s status page in. Topic tags: general, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fak
Downdetector recorded more than 2,700 reports of ChatGPT problems by 3:55 p.m. Pacific on Friday, September 25, 2026. Users also reported “Unexpected status 401: Incorrect API key” errors. OpenAI’s status checker initially said it was unaware of issues, but the company later logged a full outage affecting Codex. The records establish that users reported ChatGPT trouble and that Codex had an outage; they do not establish that every ChatGPT complaint had the same cause. 5
8
45
What happened on September 25
The reports and status updates describe two related but distinct pieces of evidence. Downdetector collects user reports, and its count showed a noticeable rise in complaints. OpenAI’s later incident record, by contrast, was specifically titled “Issues with Codex” and listed Codex Web, the Codex API, CLI and VS Code extension as affected. OpenAI marked the incident as a full outage and said the impacted services had recovered by 11:54 p.m. on the status-page record. 8
45
That distinction matters: reports of a ChatGPT problem are not, by themselves, proof that OpenAI confirmed a ChatGPT-wide outage. News coverage reported the 401 error across ChatGPT and Codex, while OpenAI’s published incident record named Codex. OpenAI did not disclose the cause of the reported errors in the available coverage. 5
45
September included several other service disruptions
The September 25 report came amid a month with other documented OpenAI incidents. On September 3, OpenAI recorded elevated errors across ChatGPT and Codex. 19 Its incident records also list elevated errors from GPT-5.6 and GPT-5.6 Instant on paid plans, as well as elevated API-model error rates later in the month.
39
42
ChatGPT Work had separate reported problems, too. OpenAI’s history includes elevated errors affecting Work across Plus, Pro, Business, Enterprise and Education plans, alongside other Work-related incidents and a problem in which mobile users could not see Work Mode and the model picker. 13
33
37
41
These records describe incidents with different dates and affected services. They should not be combined into one continuous outage or treated as evidence that every OpenAI product was unavailable throughout September.
How to read the uptime figure
The supplied OpenAI status snapshot lists API uptime at 99.96%, alongside separate figures for ChatGPT and Codex. It does not show 99.93% API uptime. 10 OpenAI also cautions that its availability metrics are aggregated across tiers, models and error types, and that an individual customer’s availability may differ depending on their tier, model and API features.
18
So an aggregate uptime percentage can coexist with a specific service or feature outage: it summarizes availability across a broader set of components, rather than guaranteeing uninterrupted access for every user. For a particular disruption, compare user-report trackers with the incident details on OpenAI’s status page—and check which product or component OpenAI actually identifies.
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "ChatGPT Outage Reports on Sept. 25: What Downdetector Saw and OpenAI Confirmed"?
On Sept. 25, Downdetector logged more than 2,700 ChatGPT problem reports by 3:55 p.m.
What are the key points to validate first?
On Sept. 25, Downdetector logged more than 2,700 ChatGPT problem reports by 3:55 p.m. September’s status history also lists a Sept. 3 ChatGPT and Codex disruption and later issues involving paid plan models, API models and ChatGPT Work.
What should I do next in practice?
The supplied OpenAI status snapshot lists 99.96% API uptime, not 99.93%; aggregate uptime does not mean every feature or user was unaffected.