Four AI Models Down Simultaneously, Fixed Within 30 Minutes

Why this matters

Four separate companies, four separate infrastructure stacks, all wobbling within the same few hours. That's not a coincidence worth shrugging off. Anthropic alone filed multiple incident reports Thursday morning, covering Claude Mythos 5.1, Claude Fable 5.1, Claude Opus 5, and a distinct Sonnet 5 error, with resolution stretching to 12:16 pm.

If you're building a product on any single model provider, this is your reminder that "the cloud" is still a handful of data centers and dependency chains that can hiccup at the same time for reasons nobody's fully explained yet. Thirty minutes of downtime sounds trivial until it's your checkout flow, your support bot, or your research pipeline stalling during business hours. Founders who assumed multi-vendor redundancy would insulate them from exactly this kind of event got a data point suggesting otherwise.

Anthropic's public incident logs deserve credit for transparency, timestamps and all, but transparency after the fact doesn't help the developer whose app silently failed at 9:23 am. Worth watching whether OpenAI, xAI, and Google publish comparably detailed postmortems, and whether anyone connects the dots on a shared root cause.