vLLM

v0.32.0Open sourceReleased 18 Oct 2026

High-throughput LLM serving engine for Linux GPU servers. By vLLM project

Linux

Official sourceNo bundles

By the ToolsBharat teamUpdated 18 Oct 2026How we check

Quick answer

vLLM is open source software for Linux from vLLM project. High-throughput LLM serving engine for Linux GPU servers. Download it only from docs.vllm.ai. The latest version we track is 0.32.0 (18 Oct 2026). ToolsBharat links straight to the publisher's own download page, with no bundled offers.

Latest version:: 0.32.0 (18 Oct 2026)
Publisher:: vLLM project
Official site:: docs.vllm.ai
Licence:: Open source — Apache-2.0
Requirements:: Linux (Windows only via WSL; experimental CPU support on macOS Sonoma+ when built from source)

Our verdict

vLLM is a server-side inference engine built to serve many requests at once on data-centre or workstation GPUs, using PagedAttention and continuous batching. It is a Python library you install with pip or uv and run as an OpenAI-compatible server; it is not a desktop app. Its docs list Linux as the supported OS and say Windows is not supported natively, so Windows users need WSL.

Key features

PagedAttention and continuous batching for high throughput
OpenAI-compatible API server, plus Anthropic Messages API support
Quantisation formats including FP8, INT4/INT8, GPTQ, AWQ and GGUF
Tensor, pipeline and expert parallelism across multiple GPUs

Pros

Pro: Built for serving many users efficiently
Pro: Supports NVIDIA, AMD and Intel GPUs
Pro: Large, active project

Cons

Con: No native Windows support (WSL needed)
Con: Command-line and Python only, no GUI
Con: Apple silicon CPU support is experimental and needs a source build

What's new in 0.32.0

See what changed in 0.32.0 in the official release notes

Version history

Version 0.32.0Latest
18 Oct 2026
SHA-256
The developer does not publish one
Source
Release notes

Requirements

OS: Linux (Windows only via WSL; experimental CPU support on macOS Sonoma+ when built from source)
Architecture: 64-bit (x86_64), ARM64
Also needs: Python 3.11 to 3.14; NVIDIA GPU with compute capability 7.5 or higher (e.g. T4, RTX 20xx and newer), or AMD ROCm / Intel XPU / supported CPU

Licence

Type: Open source
Licence: Apache-2.0
Publisher: docs.vllm.ai

Where these facts come from

Checked 30 Oct 2026

FAQ

Can I run vLLM on Windows?

Not natively. The vLLM docs say Windows is not supported and suggest using WSL with a compatible Linux distribution. For a desktop chat app on Windows, a tool such as LM Studio or Ollama is a simpler fit.

User reviews

User reviewsयूज़र रिव्यू

No reviews yet

Used vLLM? Reviews appear after moderation; we never import or invent ratings.

Alternatives