OpenAI's ChatGPT service experienced a partial outage this afternoon, disrupting access for numerous users. The company reported elevated error rates, which were resolved by 5:14 PM ET. A related issue with the API's fine-tuning component remains under monitoring after mitigation efforts.
On February 3, 2026, ChatGPT faced a partial outage that impacted many users, leading to difficulties in accessing the AI chatbot. Reports on DownDetector surged to more than 12,000 at the peak of the issue, highlighting the scale of the disruption.
OpenAI quickly acknowledged the problem, issuing a status update about "elevated error rates" affecting both ChatGPT and its Platform users. The core outage was marked as resolved at 5:14 PM ET, restoring normal service for most.
However, an active alert persists for the fine-tuning component of OpenAI's API service. The company's latest statement notes: "We have applied the mitigation and are monitoring the recovering."
This incident was not isolated to OpenAI. Anthropic's Claude AI chatbot also encountered a similar outage, with "Elevated error rate on API across all Claude models." That issue was resolved earlier, by 1 PM ET.
Such outages underscore the growing reliance on AI services and the challenges in maintaining uptime amid rapid scaling. OpenAI has not detailed the root cause, but the swift resolution minimized prolonged disruptions for users worldwide.
