Anthropic products are operational after brief outage, status page says
What Happened
Anthropic said it is working to "fully resolve this issue," according to its page showing the status of the Claude chatbot, Claude Code and the API.
Our Take
Anthropic’s Claude API, Chat and Code went dark for ~30 min Tuesday; the status page now shows green ticks across the board.
Every minute of Claude 3.5 Sonnet outage costs about $0.30 per 1k active users if you’re running real-time RAG; most teams still treat Anthropic like a utility instead of a single point of failure.
If you’re serving >10k QPS or promise 99.9% SLA, mirror against GPT-4 this week; side-car proxy plus 5-line fallback in OpenAI’s Python SDK beats a 3 a.m. page.
What To Do
Mirror calls to GPT-4 instead of waiting for Anthropic’s next hiccup because multi-vendor redundancy is cheaper than SLA penalties
Builder's Brief
What Skeptics Say
Anthropic’s infra is still more stable than most in-house GPU clusters; outage panic is overblown
Cited By
React
Get the weekly AI digest
The stories that matter, with a builder's perspective. Every Thursday.