AI LEADSpace

vllm-project/vllm

Inference
Rank of 1903

Known for

Serving a model to real traffic — batching, paged attention and an OpenAI-shaped endpoint.

The serving layer behind a large share of production inference. Low star count relative to its reach, because the people who depend on it are operators rather than browsers.

78.1index score · Open Source
Stars55.3 ×0.25
Forks71.1 ×0.33
Activity97.4 ×0.42

Factors are percentiles within the 19-entry ranked cohort. Weights are renormalised over the factors that have data.