vLLM (@vllm_project) — Twitter Score 30 out of 1000

vLLM is an open-source LLM inference and serving engine built around PagedAttention, an attention algorithm inspired by virtual-memory paging techniques. vLLM's Twitter Score is 30 out of 1000 as of September 2026, ranking #32,202 (top 10%) among 487,142 tracked accounts.

vLLM was founded by Woosuk Kwon, Zhuohan Li, and Simon Mo, and originated in UC Berkeley's Sky Computing Lab. Licensed under Apache 2.0, the project is designed for high-throughput, memory-efficient inference, with original research reporting throughput improvements of 2-4× at equivalent latency compared to evaluated systems.

vLLM is affiliated with the PyTorch Foundation as a foundation-hosted project, a status announced on May 7, 2025, and is also affiliated with the Linux Foundation. The project has received backing from Andreessen Horowitz, Sequoia Capital, and ZhenFund. Official documentation lists multiple corporate and cloud providers as sponsors and compute resource contributors. vLLM has 48,483 followers as of September 2026 and has published 1,275 posts. vLLM has been mentioned by 72 accounts a total of 237 times.

Who follows vLLM on X?

The accounts that follow vLLM include 170 founders, 43 angel investors, 37 influencers and 1 VCs, among them Tarun Chitra, mikedemarais.eth, Matthew Graham, Andrew Chen, Zmanian and Alex Atallah.