Why Apple’s M6 Chip Can’t Hold Its Own Against Any NVIDIA GPU

In introducing the latest Mac mini, Apple also debuted its sixth generation of M-series processors for Macs. Apple of course extols the M6 chip as extremely powerful, providing significant boosts to CPU and performance, and there’s at least a little merit to those claims. In our review testing, the machine didn’t miss a beat in everything from photo editing to 4K video transcoding.

If you’re not wedded to the Apple ecosystem, it’s important to weigh the M6’s processor tech against what’s available on the PC side of the fence. And if you’re serious about graphics, that probably means computers with a dedicated NVIDIA GPU. In fact, you’ll want a dedicated GPU for meaningful performance on most PCs since integrated graphics chips tend to be good for little more than streaming and light gaming.

In short, most NVIDIA GPUs are going to wipe the floor with the M6 when it comes to gaming and other real-time graphics tasks. That said, if you’re willing to lock into the Apple ecosystem, the M6 can potentially deliver what you need while costing a lot less.

It’s a bit difficult to offer an apples-to-apples comparison, if you’ll forgive the joke. Whereas the M6 is an all-in-one processor with 12 CPU and 12 GPU cores, the performance of NVIDIA GPUs is dependent not just on the card you pick but the AMD or Intel CPU you pair it with. Moreover, the best way of benchmarking Mac graphics is often through Metal scores, and Windows PCs can’t run software using the Apple-optimized Metal API. They veer toward DirectX 12 or Vulkan instead.

Thankfully, there is a roundabout way of tackling the question. MacRumors notes that the Geekbench OpenCL score for the new Mini’s other processor option, the M5 Ultra, is roughly in line with NVIDIA’s GeForce RTX 4080. The M5 Ultra also has a 366,744 Metal score. By contrast, the M6 has a Metal score of 93,217, nearly four times less. 

That’s bad because the RTX 4080 represents last-generation NVIDIA architecture from 2022. You can already push the limits of RTX 50-series GPUs in some games (speaking from personal experience), and since real-time 3D graphics are continually evolving, it’s doubtful the M6 will be able to keep pace at all by 2028 when the RTX 60-series cards are rumored to come out.

A quick reality check, however — if you’re not obsessed with maximum detail or futureproofing, the M6 does fine. We found the M6 Mini averages 55 frames per second in Cyberpunk 2077 with default Mac quality settings, which is plenty smooth, and many games aren’t as demanding. For tasks that don’t involve real-time rendering, like video transcoding, any gap with a comparable NVIDIA-based PC should be minor to non-existent. Most people probably don’t care if a video exports a few seconds faster.

The M6 wins in one major aspect, and that’s cost. Although the new Mini’s $900 sticker price is a $300 hike over the previous generation, even an RTX 5060 is going to cost you north of $450, and that’s before you pair it with the other components needed to build a working PC. Meanwhile, high-end RTX 5080 cards are going for over $2,000, which is enough to buy a MacBook Pro.

Viewed in that light, it may easily be worth putting up with Apple’s less impressive performance. You could buy an M6 Mini today, then the inevitable follow-up two years later, and still save money versus buying some NVIDIA-powered gaming laptops. (The Mini doesn’t ship with a display or input devices, mind you.)

One problem with all Macs is compatibility. While there’s no shortage of productivity apps for macOS, most games are PC-first not just because of an existing customer base, but also that greater amount of horsepower. If I were to choose a computer primarily for gaming, I’d rather buy Valve’s $1,050 Steam Machine, which can run most modern games with console-like simplicity despite being based on custom AMD hardware.