Primate Labs, the creators of the popular benchmarking tool Geekbench, have released a new tool called Geekbench AI, 27 that is now available on iOS, Android, Windows, macOS, and Linux. This tool measures how well your device handles real-world machine learning tasks, assessing the performance of the CPU, GPU, and NPU. The scores provided are based on both accuracy and speed, offering a clear view of your device’s AI capabilities.
Previously known as Geekbench ML during its preview phase, Geekbench AI has now been rolled out for general use. It provides three key scores—single precision, half-precision, and quantized—along with an accuracy measure to evaluate how well models perform tasks. Additionally, the tool supports a range of AI frameworks, including CoreML for macOS and iOS, OpenVINO for Windows and Linux, QNN for Snapdragon-powered Arm PCs, and other vendor-specific frameworks for Android devices. Each workload is tested multiple times to ensure devices perform at their peak.
What are your thoughts on this new tool? Do you think Geekbench AI will become the standard for measuring device performance in machine learning tasks? Share your opinions below.
