Hornet is a GPU-optimized nuclear data processing code, designed to enhance the efficiency of handling neutron cross-sections and ACE file sampling for incident neutron data and thermal scattering laws. Developed using CUDA C++ with an object-oriented approach, Hornet is independent of any specific Monte Carlo code, making it suitable for academic research and seamless integration into Monte Carlo frameworks. Verification tests conducted both independently and in collaboration with GREAPMC – an in-house GPU-focused Monte Carlo code – confirm Hornet’s accuracy, aligning closely with MCS, an in-house high-fidelity CPU-optimized Monte Carlo code. These tests include benchmark simulations such as thermal scattering and Mosteller benchmarks. Notably, Hornet achieves root mean square (RMS) power differences below 0.1% in pin-by-pin power verification of an OPR1000 fuel assembly, demonstrating its precision. Additionally, it significantly outperforms MCS in computational speed, providing a GPU-optimized solution for continuous-energy neutron transport