Four of the world's largest cloud-based AI models—operated by OpenAI, Anthropic, xAI and Google—experienced a rare and unprecedented set of overlapping service interruptions on Thursday morning. Anthropic's Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 suffered a partial outage, which was resolved by 12:16 pm. A separate issue with Claude Sonnet 5 saw a brief surge in errors shortly after noon.

Meanwhile, OpenAI reported elevated errors across ChatGPT and Codex, leading to degraded performance. A mitigation put in place a little after 11 am saw the issue marked as resolved by 12:55 pm. xAI's Grok model also faced difficulties, with a user-facing error message indicating that the company was working to restore service. Reports on DownDetector surged from less than 10 to 1,365 in just 45 minutes, before falling to 273 as of this writing.

The incident has raised eyebrows among observers, with some speculating on the implications for AI reliability and others pondering if it's a sign of growing AI self-awareness. Whatever the reason, one thing is clear: the digital world just got a little noisier.