Current status: mitigated, recovery still being monitored
The latest official update says OpenAI has mitigated the issue affecting ChatGPT conversations and is monitoring recovery . On the status page, the incident is still labeled Monitoring with degraded performance rather than Resolved.
That makes the current situation different from an active investigation. OpenAI has already applied a mitigation and believes the service is recovering, but it is keeping the incident open while it checks whether error rates remain stable.
Live incident status: Elevated errors in ChatGPT conversations — OpenAI Status .
What happened to ChatGPT conversations?
OpenAI reported elevated error rates affecting some ChatGPT conversations . In its first update, the company said error rates were already improving while the team continued assessing recovery.
The status record does not currently publish a technical root cause, and it does not identify a specific model as the cause. It would therefore be inaccurate to blame a particular ChatGPT model, subscription tier, browser, or user action based on the information OpenAI has released so far.
The safest interpretation is narrower: some conversation requests were failing at a higher-than-normal rate, OpenAI detected the problem, and the company subsequently applied a mitigation.
Official wording: OpenAI incident record .
Official incident timeline
| Status | OpenAI update |
|---|---|
| 4:58 PM · Investigating | OpenAI said it was investigating elevated error rates affecting some ChatGPT conversations. It also said error rates were improving and that the team was continuing to assess recovery. |
| 5:55 PM · Monitoring | OpenAI said it had mitigated the issue affecting ChatGPT conversations and was monitoring recovery. |
Important: these times are reproduced from the timestamps displayed on OpenAI’s incident page when this article was checked. The incident can receive additional updates after publication.
What does “mitigated” mean here?
On a live service incident, mitigated means OpenAI has taken action that reduced or contained the problem enough to move from investigation into recovery monitoring. It does not mean OpenAI has formally closed the incident.
- Investigating
- The service provider is still identifying or diagnosing the problem and its impact.
- Monitoring
- A mitigation has been applied and the provider is watching service behavior to make sure recovery holds.
- Resolved
- The provider has closed the incident after determining the affected service has recovered sufficiently.
For this incident, OpenAI is currently at the middle stage: mitigation has been applied, but recovery is still being watched.
What services are affected?
OpenAI’s incident page currently lists two affected components: ChatGPT Conversations and FedRAMP. It does not list every ChatGPT feature or every OpenAI API component as affected by this specific incident.
| Component | What the incident record confirms |
|---|---|
| ChatGPT Conversations | The incident directly concerns elevated error rates affecting some ChatGPT conversation requests. |
| FedRAMP | FedRAMP is also listed as an affected component on the incident page. |
| Other ChatGPT features | This incident record does not establish that every other ChatGPT feature is affected. |
| OpenAI API | The incident page does not currently list the general API as an affected component for this specific event. |
OpenAI also cautions on its status site that availability metrics are reported in aggregate and that individual customer availability can vary depending on subscription tier, model, and features in use.
Affected components and availability note: OpenAI Status .
What should you do if your ChatGPT conversation still errors?
Because OpenAI has already acknowledged and mitigated a service-side incident, repeated local troubleshooting may not be necessary while recovery is still being monitored. A few low-risk checks can help without risking your work.
-
Check the live incident page first. If the status is still Monitoring, your failure may still be part of the same recovery period.
-
Retry the failed turn once. If the service is recovering, a later request may succeed without any account change.
-
Preserve unsent or important text. Copy long prompts or work-in-progress text before refreshing or changing chats so a service error does not cost you your draft.
-
Test another conversation only if necessary. If one thread repeatedly fails while a new conversation works, that can help distinguish a thread-specific problem from a broader service issue.
-
Avoid destructive troubleshooting. There is no evidence in OpenAI’s incident report that deleting conversations, clearing your account history, or changing subscriptions is required to recover from this incident.
Has OpenAI fully resolved the conversation error incident?
Not yet, according to the status shown when this article was checked. The current state is Monitoring, and OpenAI says it is monitoring recovery after mitigation.
The distinction matters for search results and social posts. Saying “ChatGPT is fully recovered” would go beyond OpenAI’s current wording. The more accurate statement is that the issue has been mitigated and recovery is being monitored.
If OpenAI later changes the incident to Resolved, the headline and first paragraph should be updated so the article does not continue presenting a monitoring state after the incident has closed.
Current takeaway: OpenAI has moved beyond investigation, but the incident remains open. Users may see improving reliability while OpenAI confirms that the mitigation holds.
Latest status: OpenAI Status incident page .