Over the past year, the only thing that has been talked about in the technology sector is artificial intelligence. For Google, it has become an obsession, implementing Gemini and other language models in all its applications and implementing new functions based on machine learning. It is not for nothing th at the new Pixel 9 has been presented primarily as a mobile for AI.
It is doubtful that there is really a demand for so much artificial intelligence; but there is no doubt that we are at the beginning of this technology and that the possibilities are enormous. We already see it in applications like Google Photos, capable of removing and putting people in our photos, or platforms like Galaxy AI, capable of translating conversations in real time.
However, most of these AI-based functions have a negative aspect: they require an Internet connection because in reality, they do not work on our mobiles, but on the servers of companies. Experts agree that AI will really gain popularity when mobile phones are able to run it without help; and little by little we are getting to this situation, with PC computers and Copilot+ mobile phones equipped with processors with NPU (neural processing unit).
But that brings us to the question: How capable is my phone of running AI? Are you ready for the future? To answer that question, Geekbench creators Primate Labs come up with their new app. Geekbench AI is a version of its popular benchmarking app designed specifically to test a processor’s performance in machine learning tasks.
Geekbench is one of the most popular applications to check the power of a mobile phone, tablet or computer; Its tests consist of a series of tests that simulate real w ork on tasks using one or more cores, covering a variety of scenarios. Similarly, Geekbench AI performs tests that cover different levels of accuracy, since not all features based on machine learning are created equal.
After installing the application and running the “benchmark”, Geekbench AI will show us different scores, simple precision, average precision and quantification. For the average user, all this can be a bit complicated, but the important thing is that, the higher the final score the better it should be about performance in applications that use AI locally.
In particular, Geekbench AI should be particularly useful for benchmarking mobile devices and the APIs used by developers; for this purpose, Geekbench AI is available for Android, iOS, Windows, and Mac. Although, as usual with benchmarks, there is no substitute for real-world device usage.