route-switch investigates the cost-quality frontier in LLM inference. When should you route a query to a small, fast model versus a large, expensive one? This project implements MIPROv2-based automatic prompt tuning and model selection to answer that question empirically.
Cost-quality LLM router that uses MIPROv2 to auto-tune per-model prompts against live traffic.
route-switch is one option in a category that includes LiteLLM, Portkey, OpenRouter, Unify , and LLM gateways. Our Compare page has the full side-by-side.