Apple M5 Max performance prediction: Geekbench score may exceed 250,000 mark

[ Gearbest Technology News]Recently, foreign media reported that Apple will release MacBook Pro models equipped with M5 Pro and M5 Max chips within two weeks. Based on calculations based on the performance gap of the existing M4 series chips, M5 Max is expected to bring “amazing” performance improvements, especially in terms of graphics processing, which may set a new record.

Macworld's Jason Cross compared the performance gap between the standard M4 and M4 Pro/Max, and applied it to the basic M5 for extrapolation, and came up with the following estimated data:

The M5 Max's CPU single-core Geekbench 6 score is expected to be about 4,500 points, and the multi-core score may exceed 31,000 points. The performance level is comparable to the top 64-core AMD Threadripper processor. In terms of GPU, the M5 Max equipped with a 40-core GPU is expected to become Apple's first chip to break 250,000 points in the Geekbench 6 computing test. In comparison, the M3 Ultra with its 80-core GPU scored slightly below this threshold. If Apple can achieve this breakthrough with half the cores within two generations of chips, it will mean a significant improvement in energy efficiency. In terms of graphics performance benchmarking, in the Steel Nomad test, the M5 Max is expected to score more than 4,600 points, with performance comparable to the laptop version of the NVIDIA GeForce RTX 4070, providing strong support for basic games and professional applications.

In addition to performance improvements, the M5 Pro and Max chips may also see changes in configuration. Well-known Apple analyst Ming-Chi Kuo pointed out that Apple will use TSMC’s latest SoIC-mH (system-level integrated chip-horizontal plastic packaging) packaging technology. This technology may allow Apple to more freely match the number of CPU and GPU cores. While full customization is not expected, users may be able to choose an M5 Pro or Max configuration with more GPU cores and fewer CPU cores to better match the needs of specific workloads.

magicCubeFunc.write_ad(“dingcai_top_0”);