vLLM Semantic Router introduces System Duo for routing fast and deep reasoning

vLLM Semantic Router introduces System Duo, which uses fast decision models for quick calls such as triage, guard, verify and check, according to a vLLM post on X. The router sends deep reasoning tasks to a larger model only when needed.

The project says System Duo is open source and runs on vLLM. The announcement credits XunzhuoLiu and the vLLM Semantic Router team.