GPUStack
An open-source GPU cluster manager that deploys models on vLLM, SGLang, TensorRT-LLM or other engines and serves them through OpenAI- and Anthropic-compatible APIs.
About GPUStack
Teams running their own GPUs have to pick inference engines, schedule models across mixed hardware and expose them as a governed API.
Who it’s for
- Platform teams serving models on their own GPU servers or Kubernetes
- Organisations offering internal model-as-a-service
When to consider it
Consider GPUStack when you want to run models across on-premises, Kubernetes and cloud GPUs from one control plane, including non-NVIDIA accelerators such as AMD and Ascend.
Tradeoffs & limitations
- Apache-2.0 licensed and self-hosted; GPU worker nodes must run Linux (macOS is not supported for workers), and an Enterprise edition with control-plane high availability and multi-tenancy is sold through sales.
SOTA overview · Documentation-based assessment · Sources & review method
Updates
No updates shared yet.
Discussion
Newest firstAsk a question or share how you use GPUStack.
Keep it helpful. Community rules
Loading discussion…
Sources & review method
Documentation-based assessment · Oct 31, 2026 · Prepared with AI assistance; not a hands-on benchmark.
Checked by SOTA · AI-assisted documentation review. Selection advice is our assessment; verify current requirements for your deployment.
Import history & original evidence
GPUStack official website
An open-source GPU cluster manager that deploys models on vLLM, SGLang, TensorRT-LLM or other engines and serves them through OpenAI- and Anthropic-compatible APIs.
Based on official pages and announcements checked on 2026-10-31. No hands-on test, performance benchmark or popularity ranking is claimed.