# PyTorch 2.0 with CUDA 11.7 — PyTorch release, CUDA build and NVIDIA driver pairings

For PyTorch 2.0 with CUDA 11.7, pytorch release is 2.0; cuda build is CUDA 11.7; python versions is >=3.8, <=3.11; cudnn version is 8.5.0.96; build script: base architecture list is TORCH_CUDA_ARCH_LIST="3.7;5.0;6.0;7.0", recorded from its source on 2026-09-16.

- **Pairing:** PyTorch 2.0 with CUDA 11.7 _(our reading, not quoted from the source)_
- **PyTorch release:** 2.0 _(verified: appears in the quote below)_
- **CUDA build:** CUDA 11.7 _(verified: appears in the quote below)_
- **Python versions:** >=3.8, <=3.11 _(verified: appears in the quote below)_
- **cuDNN version:** 8.5.0.96 _(verified: appears in the quote below)_
- **Build script: base architecture list:** TORCH_CUDA_ARCH_LIST="3.7;5.0;6.0;7.0" _(verified: appears in its own passage below)_
- **Build script: this CUDA build's architecture line:** TORCH_CUDA_ARCH_LIST="${TORCH_CUDA_ARCH_LIST};7.5;8.0;8.6" _(verified: appears in its own passage below)_
- **GPU architectures compiled in, Windows wheel:** 3.7+PTX;5.0;6.0;6.1;7.0;7.5;8.0;8.6 _(verified: appears in its own passage below)_
- **Minimum driver (minor version compatibility):** >= 450 _(verified: appears in its own passage below)_
- **Toolkit driver, Linux x86_64:** >=515.43.04 _(verified: appears in its own passage below)_
- **Toolkit driver, Windows x86_64:** >=516.01 _(verified: appears in its own passage below)_
- **Wheel index URL:** https://download.pytorch.org/whl/cu117 _(our reading, not quoted from the source)_

## What the source says

> | 2.0 | >=3.8, <=3.11 | C++14 | CUDA 11.7, CUDNN 8.5.0.96 | CUDA 11.8, CUDNN 8.7.0.84 | ROCm 5.4 |

## Where each value comes from

This source states these in separate places, so each value is shown with the passage that states it.

### Build script: base architecture list

> TORCH_CUDA_ARCH_LIST="3.7;5.0;6.0;7.0"

### Build script: this CUDA build's architecture line

> 11.[67]) TORCH_CUDA_ARCH_LIST="${TORCH_CUDA_ARCH_LIST};7.5;8.0;8.6"

### GPU architectures compiled in, Windows wheel

> IF "%BUILD_VISION%" == "" ( set TORCH_CUDA_ARCH_LIST=3.7+PTX;5.0;6.0;6.1;7.0;7.5;8.0;8.6

### Minimum driver (minor version compatibility)

> 11.x | >= 450 | < 525

### Toolkit driver, Linux x86_64

> CUDA 11.7 GA | >=515.43.04 | >=516.01

### Toolkit driver, Windows x86_64

> CUDA 11.7 GA | >=515.43.04 | >=516.01

## Source

- https://raw.githubusercontent.com/pytorch/pytorch/main/RELEASE.md
- https://raw.githubusercontent.com/pytorch/builder/release/2.0/manywheel/build_cuda.sh
- https://raw.githubusercontent.com/pytorch/builder/release/2.0/windows/cuda117.bat
- https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html

Last verified: 2026-09-16. Review by: 2026-12-15.
Part of [PyTorch release, CUDA build and NVIDIA driver pairings](https://referencesource.org/gpu-cuda-pytorch-compatibility/).