vLLM
A high-throughput and memory-efficient inference and serving engine for LLMs open source
Where it ranks
- Self-host LLM serving at scale: Pick (90.7/100)
Pricing
Open source; you pay for your own GPUs.
Source: https://github.com/vllm-project/vllm, checked 2026-10-19
Public signals
| GitHub stars | 94.2ksource |
|---|---|
| Forks | 23.6k |
| Last push | Oct 19, 2026 |
| 7-day star change | unknown (needs 7 days of snapshots) |
| Latest stable release | v0.32.0 Oct 18, 2026 |
| Hacker News, last 30 days | 3 stories, 32 points · top: Hyper Optimizing Kimi K3 in vLLMquery |
Matched by story title “vllm”. Title matches can include unrelated stories.
Recent releases
- v0.32.0 Oct 18, 2026
v0.32.0 Highlights This release features 814 commits from 363 contributors (121 new)! Structured decisions and generic parsers : the new POST /v1/systemone endpoint answers choice , noul and score questions over a structured state ( #59299 ), is served by default ( #60677 , #6069 - proto-v0.6.0 Oct 14, 2026
vllm-proto 0.6.0 - proto-v0.5.0 Oct 9, 2026
Release vllm-proto 0.5.0