Washington: Primate Labs, the creators behind the popular Geekbench benchmarking suite, have introduced a cutting-edge tool to assess machine learning capabilities: Geekbench AI.
Now available for iOS, Android, Windows, macOS, and Linux, this new benchmark is designed to evaluate how well devices handle real-world AI tasks, confirmed GSM Arena.
Previously known as Geekbench ML during its preview phase, Geekbench AI is set to provide a comprehensive look at device performance in the realm of artificial intelligence.
The tool evaluates the efficiency and accuracy of CPUs, GPUs, and NPUs (Neural Processing Units) by measuring their ability to perform machine-learning tasks.
Geekbench AI delivers three distinct scores for each test: single precision, half-precision, and quantized.
These scores reflect not only the speed at which a device can handle AI workloads but also its accuracy.
The benchmark also includes efficiency comparisons over time, providing users with a detailed picture of their device's AI capabilities, according to GSM Arena.
Supporting a wide array of AI frameworks, Geekbench AI is versatile across different platforms. It integrates CoreML for macOS and iOS, OpenVINO for Windows and Linux, QNN for Snapdragon-powered Arm PCs, and various vendor-specific frameworks for Android devices.
Each test runs through at least five iterations to ensure that devices are tested at their peak performance.
The new tool is integrated into the Geekbench browser, allowing for seamless cross-comparisons between different devices.
This feature enables users to measure and compare the AI processing power of their devices against others, making Geekbench AI a valuable resource for both consumers and developers alike.
Geekbench AI could provide an essential tool for assessing the growing field of machine learning and artificial intelligence across a diverse range of devices. (with Agency inputs)
