Intel recently entered the discrete GPU market with their announcement of the Intel® Data Center GPU Max 1550. This GPU is the basis of the exascale supercomputer Aurora, which is composed of 63,744 Intel® Data Center GPU Max 1550s, and has been in production since January 2025. Intel significantly revised its compilers and supported programming models to target the new GPUs, such as OpenMP. This paper evaluates the standard compliance and performance of Intel’s OpenMP implementation on Aurora against other vendors’ implementations on their GPUs (MI300A, MI300X, H100, A100). We introduce OvO, a collection of OpenMP offloading tests for C++ and Fortran to assist in compliance analysis. The main takeaway is that Intel’s OpenMP implementation provides standard compliance either on a par with or higher than those of other vendors, particularly Fortran, with competitive performance for higher energy consumption.