Geekbench AI benchmark has been released, finally giving us a tool to test “AI PC” claims

There is now a benchmark for testing AI workloads. Developed by Primate Labs, Geekbench Artificial Intelligence It’s an additional component that’s now added to the more well-known Geekbench software. But now it has a simple solution to check your device’s readiness for AI. Building on its other ML (machine learning) offering, it adapted it into something that’s more recognizable and is now booming with the AI ​​naming scheme.

With such a wide range of hardware capable of running these tasks, Geekbench AI is quite flexible in how it can be run. With support for macOS, Windows, Linux, Android, and iOS, you can check virtually any device. Along with any capable components on board, you can choose the back end to target. UPC, GPUo NPU with category separation to facilitate comparisons and testing.

It should also help to review and test some of those PC Copilot+ that require an onboard NPU with a certain amount of power. Along with any hardware that claims the same without it, something like We analyze the MSI Prestige 14 AI Evo And I wondered how well a CPU-only laptop can perform on any kind of AI workload. So it's certainly a welcome addition to our test suite.

Geekbench AI Leaderboards and Features

One of the highlights of having a good simple benchmark is being able to compare your hardware to others. For example, in Cinebench or 3DMark, you can easily see how yours compares to others, or if you're thinking of buying one, see which one ranks best for the job. In that case, Geekbench simply gives you a score for your benchmark performance so you're on the shortlist. leaderboard.

There you can see all the results and recent additions across the board, where you can compare models like the 6900XT and the Samsung Galaxy S22 Ultra, as seen in the screenshot below. After a while, these should move to the ML Benchmarks Table where hardware is classified in relation to each other.

There is also a PDF of Geekbench AI workloads Uses you can dig into. These are much more specific to what AI is being used for and not a general use of the hardware. As we see, computer vision and processing workloads are included. Image classification, face detection, and object detection are in addition to text classification. These are all real-world scenarios for the technology and are a good way to test it out.