Hey, thanks for the detailed report with Request IDs and stack traces.
This is a known issue on our side. One of our model providers is having an outage, and it’s affecting multiple models at once. It’s not related to your network or setup.
The team is aware. For now, as a workaround, if a specific model errors out, try switching to a different model and resend the request. Based on our server logs, some models are more stable than others right now.
I’ll post an update here when we have more info. Sorry for the trouble.