vLLM Semantic Router, KR Labs and AMD release Vela 3.0 open model family

vLLM Semantic Router, KR Labs and AMD released Vela 3.0, a family of six open models from 0.3B to 27B parameters that return routing decisions and extracted spans in one call, according to a post by vLLM (@vllm_project).

The collection is available on Hugging Face.

The vLLM post says the 380M model ranks first on the Decision Index among models up to 500M parameters and beats GPT-5-mini at zero-shot classification. These are the project's claims; the post does not include benchmark details.