Hey, thanks for coming back with an update, glad it cleared up.
About the repeated slowdowns, Auto routes between multiple models, and sometimes specific providers or routes on our side start lagging, which causes the Taking longer than expected message. It usually fixes itself after a few minutes, like you saw.
If it happens again, send the Request ID here plus an approximate timestamp with your timezone when it was really bad. That helps a lot to match it with our backend metrics. As a temporary workaround, you can manually switch to a specific model Sonnet or GPT-5, it often helps bypass a stuck route.