nvidia/cuopt repository overview

⁠NVIDIA cuOpt

⁠What is NVIDIA cuOpt?

NVIDIA® cuOpt™ is a GPU-accelerated optimization library that solves Mixed Integer Linear Programming (MILP)⁠, Linear Programming (LP)⁠, and Vehicle Routing Problems (VRP)⁠. It enables solutions for large-scale problems with millions of variables and constraints, offering seamless deployment across hybrid and multi-cloud environments.

Using accelerated computing, NVIDIA® cuOpt optimizes operations research and logistics by enabling better, faster decisions.

⁠System Requirements

Please refer to system requirements for NVIDIA cuOpt⁠.

⁠Image Tag Naming Scheme

The tag naming scheme for NVIDIA cuOpt images incorporates key platform details into the tag as shown below:

<image_name>:<version>-<cuda_version>-<py_version>

For example, the tag cuopt:25.8-cuda12.8-py3.12 indicates the following:

25.8-cuda12.8-py3.12
 ^        ^      ^
 |        |      Python version
 |        |
 |        CUDA version
 |
 cuOpt version

latest tag will always point to latest release and will be overwritten in every release.

⁠Usage

This should open a shell to the container. From here, you can run the cuOpt CLI, Python API, and more.

docker run -it --gpus all --rm -v $(pwd):/workspace -w /workspace nvidia/cuopt:25.8.0-cuda12.8-py312 /bin/bash

If you want to run cuOpt server, you can use the following command:

docker run -it --gpus all --rm -v $(pwd):/workspace -w /workspace -e CUOPT_SERVER_PORT=8000 -p 8000:8000 nvidia/cuopt:25.8.0-cuda12.8-py312

⁠More Information

• NVIDIA cuOpt docs⁠
• NVIDIA cuOpt GitHub⁠
• NVIDIA cuOpt Examples⁠