Apple M6 Mac mini Posts 4,610 Geekbench Score, Trailing A20 Pro in Single-Core

Q: Why does Apple's next-gen 2nm desktop chip actually lose to an iPhone in single-core speed?

Apple’s M6 does not appear to be “losing” in any meaningful desktop sense; the early data suggests Apple tuned it for sustained throughput, not maximum peak single-core speed. The key evidence is the benchmark itself. The leaked Mac mini M6 scores about 4,610 single-core and 20,676 multi-core in Geekbench 6.3, while the iPhone 18 Pro’s A20 Pro posts roughly 4,725 single-core and 12,575 multi-core. That means the A20 Pro is only about 2.5% ahead in single-core, but the M6 is far ahead in multi-core because it has a 12-core 2+4+6 layout and more thermal headroom. The most likely reason is clocking and core design choices. The A20 Pro reportedly peaks at 4.93GHz, around 3% higher than the M6 sample. A small clock edge is enough to explain a small single-core lead in Geekbench. An unverified but plausible leak says the M6 keeps a wider 10-wide decode “Super” core, while the A20 Pro uses a narrower 9-wide version that may clock higher. If true, Apple split the A-series and M-series designs to optimize for different goals: the A20 Pro for peak burst speed, the M6 for broader desktop balance. There are also reasons to be cautious. This M6 result appears to come from an engineering sample running a pre-release macOS 27 build, so clocks, scheduling, and firmware may not be final. And process leadership alone does not guarantee higher single-core scores; architecture, voltage, timing margins, and product tuning matter more than the “2nm” label by itself.

3 articles · Updated · Notebookcheck.net · Sep 15

A Geekbench 6.3 listing for the Mac mini M6 showed 4,610 single-core and 20,676 multi-core, putting Apple’s new desktop chip about 2.5% behind the A20 Pro’s 4,725 in single-core performance.
The gap appears tied to clocks and core design: the A20 Pro peaks at 4.93 GHz, roughly 3% above the M6, while an unverified leak says M6 kept a 10-wide decode core and A20 Pro adopted a 9-wide design tuned for higher speeds.
Against the outgoing 10-core M5, the listed M6 result still implies roughly a 10% single-core gain and a 20% multi-core gain, broadly matching Apple’s claim of up to 1.2x faster multi-threaded performance.
The A20 Pro remains far behind in multi-core at 12,575 because of lower core count and tighter thermal limits, while the Mac mini entry shows a 2+4+6 CPU layout and 32 GB memory.
Because the benchmark likely comes from an engineering sample running macOS 27.0, final retail M6 performance could shift once broader testing appears.