Interactive runs are paused while we bring dedicated capacity online.

Lewn needs a GPU that stays awake. Ours spins up on demand right now, so a single interactive answer takes about five minutes.

It is off for everyone on every plan until that capacity is live. Nothing about your account is limiting you here.

Queued studies are running normally today. You send the question and the audience, we field it on the calibrated panel, and the report comes back. It is the same model and the same calibration that sits behind Lewn.