OpenAI's generative artificial intelligence service ChatGPT suffered access failures for some users. Anthropic's rival service Claude drew a wave of outage reports the same day, reviving questions about the reliability of major generative AI services.
Nearly 1,000 outage reports related to OpenAI were filed with the tracking site Downdetector around 2 p.m. Eastern time on Sept. 29, Forbes reported. About 85% of the reported problems involved ChatGPT use.
OpenAI said on its official status page that ChatGPT was experiencing higher-than-usual error rates, and that some users might encounter failed requests, problems logging in or signing up, and tasks that do not run to completion. Errors were also reported in the coding agent Codex and parts of the API.
The scale of the disruption swung sharply by the hour. Downdetector reports fell below 100 around 2:30 p.m. before climbing back above 500 roughly 15 minutes later, according to Forbes. That pattern suggests intermittent problems recurring in certain regions and services rather than a full outage locking out all users at once.
The disruption came as OpenAI held its annual developer conference the same day. At the event, the company unveiled Dots, a new AI agent that carries out tasks continuously on a user's behalf. Built on GPT-6 Astra, Dots runs on its own cloud computer and is designed to perform long-running tasks using applications the user has connected.
Anthropic's AI chatbot Claude also went down the same day. Outage reports for Claude on Downdetector topped 11,000 around 10:15 a.m. Anthropic said errors had hit the Claude web service, Claude Code and its API, then carried out recovery work and said service had returned to normal by the afternoon.
ChatGPT and Claude also became unstable at the same time in early September, drawing tens of thousands of user reports. In August, OpenAI's Work agent experienced elevated error rates and delayed responses for about three hours.
OpenAI recently halted the release of its next-generation model, GPT-6.1 Astra, citing AI safety concerns. Separately, the company is pressing ahead with its expansion into AI agents, including the unveiling of Dots at the developer event.
