Geekbench is known for its popular CPU and GPU benchmarks but it lacked one important benchmark, and that’s AI benchmarking. Several manufacturers have started including NPUs on their SoCs, thanks to the rise of AI in recent times.
Now, the platform has released a cross-platform benchmark to assess the AI capabilities of a machine. The platform has been working on an AI benchmark since about 2021. At the time, it was called Geekbench ML but now it has been changed to Geekbench AI, for all obvious reasons.
READ MORE: NVIDIA GeForce RTX 3070 Laptops Price List 2024
Similar to other Geekbench benchmarks, the AI benchmark is also available for Android, iOS, Linux, macOS, and Windows. You can try it by going to geekbench.com/ai/download and installing the benchmark for your device.
Upon running the Geekbench AI benchmark, you get three scores—High Precision Score, Half Precision Score, and Quantized Score. These scores, according to Labs, are supposed to evaluate “how accurately that model can do what it’s supposed to do.” In simple words, a higher precision score means better, faster, and more accurate AI performance.
READ MORE: Samsung 5G Mobile Phones Price List in India (Aug 2024)
Unfortunately, there is not much data to evaluate how this benchmark actually works and how the scores translate into real-world AI performance. Oh, and the benchmark uses several frameworks like ONNX, CoreML, TensorFlow Lite, and OpenVINO for the score assessment.
READ MORE: Automatic Bikes Price List in India 2024
You can follow Smartprix on Twitter, Facebook, Instagram, and Google News. Visit smartprix.com for the latest tech and auto news, reviews, and guides.