NVIDIA cuOpt

NVIDIA cuOpt is an open-source, GPU-accelerated engine for decision optimization designed to handle large-scale problems with millions of variables and constraints.

cuOpt supports linear programming, vehicle routing problems, and quadratic programming, with beta support for mixed-integer programming, QCQP, and SOCP.

Yes, cuOpt is an open-source engine available for developers on GitHub, PIP, Docker, Conda, and NVIDIA NGC.

cuOpt uses GPU acceleration to speed up optimization workflows, providing significant speedups for LP when lower-accuracy solutions are acceptable and strong routing performance for VRP.

Yes, enterprise support is available for production deployments through NVIDIA AI Enterprise, which provides security, reliability, and enterprise-class support.

cuOpt supports minimal-code integration with modeling languages including AMPL, CVXPY, PuLP, GAMSPy, and JuMP, and extends into agent-first workflows through open-source cuOpt agent skills.

Yes, cuOpt supports dynamic and batch optimization, allowing users to continuously adapt to changing variables and constraints by rerunning models in near real time.

cuOpt is available as open-source software on GitHub and can also be accessed via PIP, Docker, Conda, and NVIDIA NGC.