The Procyon AI benchmark measures inference times on popular models like MobileNet, ResNet, and YOLO. It's designed to show how well a system can accelerate machine learning tasks in real use.
The ACEMAGIC K1 posts an overall UL Procyon AI score of 76, which places it near other Iris Xe based systems in the chart. Lighter models like MobileNet V3 complete in about 1.73 ms, while heavier workloads such as ResNet 50 and YOLO V3 take 17.55 ms and 128.81 ms respectively. Results are perfectly usable for basic AI-assisted features in modern apps, though newer Mini-PCs with RDNA 3 or Arc graphics finish these workloads noticeably faster.