The response time to a request has reached 4-5 minutes. People forget what they had done. It is no longer possible to work fluently. What is the reason for this slowness?
Well. It’s almost always Anthropic’s inference API. Timeouts, empty responses, garbled responses — outages galore. They don’t publish all those “minor implications” on their status website.
So many outages. My requests never come back, out for hours, all models. Second day in a row multiple hours of outages. Anyone else have 4+ hours outage today 5-9pm PST?
For those using Claude, see frequent outages at Anthropic: https://status.anthropic.com/