Model Routing, Slowdowns & Rate Limits
Sep 26, 2026 · 457 messages · 56 active members
Luna dropped from 120-150 TPS to 35-40 while Baseten held at 33k tokens/sec. Omniroute was recommended as a Codex account-rotation wrapper despite long-call bugs, and @robinroy sparked interest in auto-rotating routers a…
Read full digest →