Apple recently officially announced the new Mac Studio and Mac mini equipped with the newly released M6 and M5 Ultra chips, marking a historic leap in desktop hardware and local artificial intelligence computing power.

As the core highlight of this product lineup, the M6 is Apple's first self-developed chip using 2-nanometer process technology. In terms of architecture design, it features a more powerful CPU with 12 cores and a 12-core GPU, and for the first time, it introduces two 16-core neural engines (Neural Engine). According to official data and test results, when running large language models (LLMs) in LM Studio, the Mac mini equipped with the M6 chip demonstrates terrifying performance superiority in prompt processing speed, showing a 13.5 times increase compared to the early M1 version, and a 4.8 times leap compared to the previous generation M4 version. In addition, the Mac mini (M5 Pro) presented at the same time also performs impressively, with LLM prompt processing performance 8.5 times faster than M2 Pro and 4 times faster than M4 Pro.

For professional and heavy AI workloads, Apple has introduced the most powerful chip in its history — M5 Ultra. This chip uses the next-generation UltraFusion technology to achieve a "four-Die" (four-chip packaging) architecture for the first time in the M-series system-on-chip (SoC). In terms of specific configuration, the M5 Ultra can have up to 36-core CPU and 80-core GPU, and it also features a unified memory bandwidth of up to 1.2TB/s, which is 50% higher than that of the previous generation M3 Ultra.

From the trend of large model applications on the edge, the collective surge in hardware computing power not only provides powerful local reasoning support for professional developers but also allows desktop devices to completely free themselves from excessive dependence on the cloud when handling large AI tasks, marking the official entry of personal computers into a new era of edge-side large model computing power.