---
id: custatevec-cu12
version: "1.14.0"
license: NVIDIA Proprietary Software
license_treatment: unclear
maintenance: active
---

# custatevec-cu12 — cuStateVec - a component of NVIDIA cuQuantum SDK

License: unclear · Maintenance: active · Downloads: 75.6K/mo

## What it is and what it does

cuStateVec is NVIDIA's GPU-accelerated library for state vector quantum simulation operations, part of the cuQuantum SDK. It exposes C APIs directly and can also be accessed through the cuQuantum Python wrapper. The library is designed to offload quantum state vector computations to NVIDIA GPUs, enabling high-performance quantum circuit simulation on CUDA 12 hardware.

The package ships as platform-specific wheels for Linux (x86_64 and aarch64) and has no Python runtime dependencies, meaning it acts as a thin wrapper around the underlying CUDA libraries. Installation requires an NVIDIA CUDA 12 environment; the library is actively maintained and currently in use by quantum computing researchers and practitioners building simulators.

Use it for:

- Accelerate quantum circuit simulation by offloading state vector operations to NVIDIA GPUs
- Build high-performance quantum simulators that leverage CUDA 12 hardware for large state spaces
- Integrate GPU-accelerated quantum state operations into research or production quantum computing workflows
- Prototype quantum algorithms with GPU-backed state vector computation for performance-critical applications

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

cuStateVec provides GPU-accelerated state vector operations for building quantum simulators, optimized for NVIDIA CUDA 12 hardware.

Yes, if you have NVIDIA CUDA 12 hardware and are building quantum simulators. The active maintenance, zero security vulnerabilities, and GPU acceleration justify installation for quantum computing workloads. Verify NVIDIA's proprietary license terms for your use case first, and note that practical use likely requires the cuQuantum Python wrapper.

## Install

pip install custatevec-cu12
uv add custatevec-cu12
poetry add custatevec-cu12

## Installing custatevec-cu12

Before you install: Medium install friction due to platform-specific wheels (manylinux2014 x86_64 and aarch64 only); active maintenance with recent releases, but requires NVIDIA CUDA 12 runtime environment.

License in practice: Licensed under NVIDIA Proprietary Software with unclear terms; review NVIDIA's licensing documentation before use in commercial or redistributed contexts.

Quickstart:

pip install custatevec-cu12

import custatevec
# Access cuStateVec C APIs or use via cuQuantum Python wrapper

Requires NVIDIA CUDA 12 runtime and compatible GPU; runs only on x86_64 or aarch64 Linux systems.

Verify before relying:

- Exact Python version compatibility (requires_python not specified in metadata)
- Whether cuStateVec can be used standalone or requires cuQuantum Python wrapper for practical use
- Performance benchmarks or typical use-case throughput expectations

## Package facts

- License: NVIDIA Proprietary Software (unclear)
- Python support: unspecified
- Install friction: medium
- Maintenance: active
- Downloads: 75.6K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

quantum simulator state vector, gpu quantum computing, nvidia cuda quantum, quantum state vector library, high-performance quantum simulation, cuda quantum acceleration, state vector quantum operations, quantum-computing, gpu-acceleration, cuda

[View on SkillFed](https://skillfed.io/packages/custatevec-cu12) · [View on PyPI](https://pypi.org/project/custatevec-cu12/)