2 hrs ago
ChatGPT, Claude and Grok Hit by Simultaneous AI Outages
Several popular AI services stopped working properly at the same time.
ChatGPT, Claude and Grok users said websites and apps were slow, unresponsive or unable to answer questions.
Some people also could not log in or load their conversations.
Thousands of users reported problems on an outage-tracking website.
OpenAI confirmed that ChatGPT and Codex were showing more errors than usual.
Anthropic said some Claude models were also having elevated errors and that engineers were working on a fix.
Grok's operator confirmed that it was investigating performance problems.
Google Gemini and Amazon Web Services were also reported as possibly affected, but the reports were less fully confirmed.
The companies had not explained one shared cause for all of the disruptions.
ChatGPT, Claude and Grok experienced widespread service disruptions on Thursday evening, September 3, 2026.
Users reported unresponsive websites and apps, failed prompts, login problems, timeouts and internal server errors.
Downdetector recorded more than 30,000 reports for OpenAI, about 1,400 for Claude and more than 1,000 for Grok.
OpenAI said elevated errors affected ChatGPT and Codex, while Anthropic reported problems affecting several Claude models.
Reports also indicated possible disruptions involving Google Gemini and Amazon Web Services, though the articles did not provide equivalent confirmation from those services.
- Who
- Users of ChatGPT, Claude, Grok and potentially Google Gemini and Amazon Web Services; the services' operators were investigating or mitigating the problems.
- What
- Simultaneous or overlapping technical disruptions caused elevated errors, failed prompts, timeouts and unresponsive applications.
- Where
- Across the services' web, mobile and desktop platforms and, for Claude, its API and Claude Code.
- When
- Thursday evening, September 3, 2026; OpenAI said its incident had been ongoing for at least 36 minutes in one update.
- Why
- The overall cause was not publicly explained; Anthropic said it had identified a cause for elevated errors affecting specific Claude model requests.
Key facts
- Most reported outage
- OpenAI received more than 30,000 user reports on Downdetector.
- Claude reports
- Downdetector recorded approximately 1,400 reports for Claude.
- Grok reports
- Downdetector recorded more than 1,000 reports for Grok.
- OpenAI status
- OpenAI reported elevated errors across ChatGPT and Codex and said mitigation had been applied while recovery was monitored.
- Claude status
- Anthropic initially identified errors involving Mythos 5.1, Fable 5.1 and Opus 5; later, Opus 4.8 and Opus 5 remained affected.
- Reported symptoms
- Users encountered inaccessible services, failed prompts, Gateway Timeout messages, login difficulties and requests that timed out.
- Grok status
- SpaceX AI or SpaceXAI confirmed it was investigating Grok service or performance issues.
Quotes
OpenAI status page
The official status page for OpenAI services
“We have applied the mitigation and monitoring the recovery”
telegraphindia.com
CJ Avilla
Claude developer who reported an infrastructure-related partial outage affecting Claude services
“We’re working on it but don’t have an ETA yet.”
CNBC TV 18
Sources
Major AI platforms ChatGPT, Claude, and Grok hit by simultaneous service disruptions
ChatGPT, Gemini, Claude disrupted as AI outages spread, with AWS and Grok also affected
ChatGPT, Claude and Grok down? Users report widespread outages across major AI platforms






