Cloud-based AI services from multiple major providers experienced overlapping service interruptions Thursday morning, with OpenAI and Anthropic reporting elevated errors that persisted for several hours.
What Happened
Anthropic reported a partial outage affecting Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 beginning at 9:23 am Eastern time. The company identified the cause approximately 15 minutes later and deployed a fix before reporting full resolution at 12:16 pm. A separate incident involving elevated errors on requests to Claude Sonnet 5 occurred briefly just after noon. OpenAI simultaneously reported degraded performance across ChatGPT and Codex starting at 10:43 am Thursday, with a mitigation in place roughly 30 minutes later and the issue marked resolved by 12:55 pm.
Why It Matters
The overlapping nature of these outages highlights the concentration risk inherent in depending on a small number of large-scale AI providers. Developers building applications on top of these APIs had limited alternatives when multiple services degraded simultaneously, potentially affecting production systems that rely on real-time AI capabilities. The incidents underscore ongoing questions about infrastructure reliability as AI becomes embedded in business-critical workflows.
The Bottom Line
Both Anthropic and OpenAI resolved their respective issues by early afternoon Thursday. Neither company disclosed the technical root cause of the disruptions.