ChatGPT, Claude, Grok and Gemini experienced simultaneous outages within a short period. Anthropic reported a partial outage for Claude Mythos 5.1, Claude Fable 5.1 and Claude Opus 5 at 9:23; the cause was identified approximately 15 minutes later, and the issue was resolved at 12:16. Short-lived errors were also seen in Claude Sonnet 5 after midday. OpenAI announced degraded performance in ChatGPT and Codex at 10:43; the issue was resolved at 12:55. While error messages continued on Grok, DownDetector reports rose from below 10 before 9 to 1,365 at 9:45, then fell to 273. Google did not confirm problems with Gemini; reports rose from 23 to 412. StatusGator showed a possible outage for the Gemini API between 10:45 and 11:15. AWS, Microsoft Azure and Cloudflare did not report major problems. It is rare for four models to be affected during the same period. Claude’s 90-day uptime rate was listed as 99.4%, ChatGPT’s as 99.63%, and ChatGPT Codex’s as 100%.
Why it matters
The outages occurring within the same time frame show that even users who are not dependent on a single AI service may experience disruptions in their workflows. However, the absence of any major problems reported by AWS, Microsoft Azure and Cloudflare leaves open the question of whether the incidents can be explained by a single failure in shared infrastructure. Differences between company reports and independent monitoring data make it difficult to determine the scope of the problems experienced by users, particularly with Gemini and Grok. Although the uptime rates appear high, this metric alone does not reveal the operational impact of short-lived outages occurring simultaneously; for users who connect these services to their business processes, the key issue is whether usable alternatives are available in similar situations.