Geekbench is a cross-platform processing benchmarking program with a scoring system that can be divided into single-core and multi-core performance, as well as the ability to simulate the workload of real-world usage scenarios. Currently, there are Geekbench 2, Geekbench 3, Geekbench 4, and Geekbench 5, the latest version of which was Geekbench 5.5.
Today, Geekbench 6 was released for Windows, Linux, macOS, iOS, and Android platforms.
Geekbench official Primal Labs said that a lot has changed in the technology world in the past three years. Smartphone cameras can take bigger and better photos. Artificial intelligence, especially machine learning, has become ubiquitous in general-purpose and mobile applications. The number of cores in computers and mobile devices continues to grow. The way we interact with computers and mobile devices has changed dramatically.
To keep up with these advances, Geekbench 6 has been released. This latest version of Geekbench has been designed with the modern user in mind, reflecting the way users will actually use their devices in 2023.
▲ Geekbench 6 running in Windows 11
Geekbench 6 for macOS screenshot ▲
The official blog says Geekbench 6 is designed to focus on machine learning workflows, reduce the focus on single-core CPU testing, and address the issue of shared workloads for processors with large and small core architectures.
In machine learning, Geekbench 6 will “make better use of GPUs” in less CPU-intensive scenarios, with support for more machine learning acceleration instructions and more uniform GPU performance across platforms.
For multi-core benchmarking, Geekbench 6 will be able to quantify the way cores share tasks in “real workload examples”. Previously, Geekbench simply added the performance of each core together.
In addition, Geekbench 6 increases the size of test images to simulate what users will actually take with their phone cameras in 2023.
Geekbench 6 will also include a number of new tests:
- Background blur, such as during a video conference
- Photo filters, similar to those used by modern social media apps
- Object detection for AI workloads
- Photo libraries for importing and semantically tagging photos and metadata
- Text processing for parsing and converting things like price drops and regular expressions in Python (a more realistic developer use case)
The following is a comparison of Samsung Galaxy S22 Ultra scores in Geekbench 6 and Geekbench 5.5.1.
Geekbench 6 is also different, no longer selling single-user licenses, but changing to free for all non-commercial use.
The Pro version of Geekbench 6 will remain at $99 (currently about RMB 675) and add more features, including offline results, an automated command line and a portable version. For the next two weeks, Primate Labs will sell the Pro version at a discounted price of $79 (currently about RMB 539).
