There’s a particular kind of Apple announcement that doesn’t come with a stage, a keynote, or a slow-motion drone shot over Apple Park. It just lands on the newsroom page one Tuesday morning, and if you weren’t paying attention you’d scroll right past it. That’s how the new Mac mini arrived on August 25, 2026: a press release, a few product shots, and a set of performance claims that are genuinely hard to process on first read.

The short version is that Apple has put its brand-new M6 chip in the little five-inch box on your desk, and for people who want more, there’s an M5 Pro option sitting above it. Apple says the M6 model is up to 4x faster at AI work, 2x faster at graphics and storage, and about 40 percent faster on the CPU than the M4 Mac mini it replaces. Pre-orders opened the same day, and the machines start showing up on doorsteps September 22.

Those numbers deserve a second look, because Apple’s generational jumps have been getting smaller, not bigger. For the last few years, the pattern was familiar: 15 percent here, 20 percent there, a new video encoder, move along. This one breaks the pattern, and the reason has less to do with the CPU and more to do with what Apple has been quietly building into its graphics cores.

The M6 is Apple’s first chip built on a 2-nanometer process, which makes it the first commercially shipping silicon anywhere to use TSMC’s newest manufacturing node. That’s a real milestone, and it’s the kind of thing Apple would have spent ten minutes of stage time on in a different year.

What it bought them is room. The M6 has a 12-core CPU, two more than the M5, arranged in a configuration Apple describes as two “super cores,” four performance cores, and six efficiency cores. The GPU also gains two cores, up to 12, and every single one of them now carries a Neural Accelerator, which is Apple’s term for dedicated matrix multiplication hardware sitting right inside the graphics pipeline. On top of that, there’s an all-new Dual 16-core Neural Engine, which the system can use both halves of simultaneously.

If that sounds like a lot of jargon, here’s the practical translation. Matrix multiplication is the single operation that dominates running an AI model, whether that’s a language model answering a question or a diffusion model generating an image. Historically, Apple’s Neural Engine handled some of that and the GPU handled the rest, awkwardly. Putting accelerators in every GPU core means the chip stops splitting the difference. Apple claims nearly 30 percent more peak GPU compute for AI than M5, and more than 8x what the M1 could do.

The benchmark Apple leads with is the one that tells the story best: up to 13.5x faster large language model prompt processing in LM Studio compared to the M1 Mac mini, and 4.8x faster than the M4. Prompt processing is the part where the model reads your input before it starts generating, and on older Macs it’s the bit that makes local AI feel like a science project rather than a tool. Cutting it by a factor of five is the difference between “this is neat” and “I actually use this.”

The less exciting numbers are worth noting too, because most people don’t spend their day running local models. Spreadsheet calculations in Microsoft Excel are up to 2.3x faster than on the M1 mini and 1.5x faster than the M4. Gaming with ray tracing in Cyberpunk 2077: Ultimate Edition runs up to twice as fast as the M4 model. Memory starts at 16GB and goes to 32GB, with bandwidth up to 170GB/s, a 10 percent bump over M5 and two and a half times what the M1 managed.

Why there’s an M5 Pro and not an M6 Pro

Here’s where the lineup gets strange, and it’s the part that will confuse anyone who assumes Apple’s naming means something. The step-up Mac mini doesn’t get an M6 Pro. It gets an M5 Pro, a chip from the previous generation.

That isn’t a supply hiccup or a staggered rollout. Bloomberg‘s Mark Gurman reported earlier this year that Apple cancelled the M6 Pro and M6 Max entirely, choosing to skip that tier and put its silicon engineering effort straight into the M7 generation instead. So for the first time in the Apple silicon era, there’s a mixed-generation Mac on sale, with the newer chip at the bottom of the range and the older architecture above it.

In fairness, the M5 Pro is no slouch. It runs up to an 18-core CPU and a 20-core GPU with third-generation ray tracing, supports 64GB of unified memory at 307GB/s, and gets Neural Accelerators in every GPU core just like the M6 does. Apple says it’s up to 8.5x faster than the M2 Pro mini at LLM prompt processing, 4.5x faster at ray-traced Blender renders, and 2.1x faster at image processing in Affinity.

It also gets the ports that matter for anyone doing serious work. The M5 Pro mini has three Thunderbolt 5 ports on the back, while the M6 model makes do with Thunderbolt 4. That gap enables something Apple mentions almost in passing but which is quietly one of the more interesting things about this machine: you can cluster multiple Mac mini systems over Thunderbolt 5 to run large AI models entirely on device. Apple has been pushing this idea on the Mac Studio side too, where a cluster of four machines delivers up to 3x faster AI inference than a single one.

Both models share the rest of the connectivity upgrades, and they’re the sort of thing that’s easy to overlook until you need them. Wi-Fi 7 and Bluetooth 6 are new. Ethernet moves up to 2.5Gb as standard, with a 10Gb option if you’re moving a lot of files around a studio. The two USB-C ports and headphone jack stay on the front, and the headphone jack now supports high-impedance headphones. There’s also genlock support over USB-C, which lets you sync a display precisely with a camera, including an iPhone 17 Pro.

Johny Srouji, Apple’s chief hardware officer, framed the whole thing around versatility rather than raw speed. “Mac mini has always been our most versatile Mac. Whether it’s being used as a home computer, powering a professional studio, or as an always-on agentic device, it’s the little Mac that can do it all,” he said in the announcement. The phrase “always-on agentic device” is doing a lot of work there. Apple is clearly imagining a world where you leave a Mac mini humming under a desk, running AI agents that handle tasks in the background, and it’s building the marketing to match.

The part nobody’s excited about

Now for the number that’s going to dominate the comment sections. The Mac mini with M6 starts at $899, or $799 with education pricing. The M5 Pro version starts at $1,699, or $1,599 for education.

To understand why that stings, you have to go back to October 2024, when Apple redesigned the Mac mini around the M4 and kept the base price at $599 with 16GB of memory. Reviewers at the time called it the best value in Apple’s entire lineup, and they weren’t wrong. It was a genuinely cheap way into a fast computer.

Then June 2026 happened. Apple raised prices across Macs, iPads, and more, and Tim Cook told The Wall Street Journal directly that memory and storage chip costs left the company no choice. The M4 mini went from $599 to $799. The M4 Pro went from $1,399 to $1,599. Now the M6 adds another $100 on top. As Macworld put it, the cheapest entry point into Apple’s desktop lineup has climbed by $300, or 50 percent, in the space of two months.

This isn’t an Apple-specific greed story, as much as it might feel like one. The entire industry is living through what people have started calling “RAM-ageddon.” Contract prices for conventional DRAM jumped roughly 90 to 98 percent in the first quarter of 2026 alone, with another 58 to 63 percent increase following in the second. The cause is straightforward: AI data centers buy memory in volumes that dwarf the consumer market and can afford to pay whatever it takes, which starves the supply that used to keep laptops and phones cheap. Apple itself said it had “never seen a component price increase this much, this quickly.“

Nobody expects this to resolve soon. Gartner projects combined DRAM and SSD prices rising around 130 percent across 2026, pushing average PC prices up about 17 percent. IDC’s Jitesh Ubrani has been blunter, saying there’s no relief in sight before the end of 2027 and that pricing is unlikely to return to 2025 levels even after capacity expands. Lenovo, Dell, HP, Acer, and Asus have all warned customers about hikes in the 15 to 20 percent range. Apple is simply the most visible company doing what everyone is doing.

So should you buy one

The honest answer depends entirely on what you were hoping for. If you wanted the $599 Mac mini to come back, it isn’t coming back, and pretending otherwise doesn’t help. That machine was a product of a moment when memory was cheap, and that moment is over for at least a couple more years.

What you get instead is a computer that’s meaningfully more capable than the one it replaces, in a way that’s rare for a mid-cycle desktop refresh. If you’re running local language models, editing video, doing 3D work, or genuinely planning to leave AI agents chugging away overnight, the jump from an M1 or M2 mini is enormous, and even the M4 comparison is substantial. The new Mac mini also arrives alongside macOS 27, which brings Siri AI and a deeper set of Apple Intelligence features into everyday apps, plus the usual round of Liquid Glass refinements. That software is in public beta now and ships free this fall.

Apple also announced a new Mac Studio the same day, with the M5 Max starting at $2,499 and a quad-die M5 Ultra configuration at $5,499, so there’s a clear ladder if the mini isn’t enough. And for the environmentally minded, the new mini is built with 50 percent recycled material overall, including fully recycled aluminum in the enclosure.

If you’re a casual user who mostly browses, writes, and watches things, the M4 mini you already own is still perfectly good, and honestly so is an M2. The upgrade math only works if AI performance is something you actually care about, which is a sentence that would have sounded ridiculous three years ago and now describes a surprising number of people.

Either way, it’s a strange sort of milestone. Apple’s smallest, most modest desktop is now a serious on-device AI machine with the world’s most advanced chip inside it, and the reason it costs more is that the AI boom made the memory inside it expensive. There’s something almost circular about that, and it’s probably worth sitting with for a moment before you hit the pre-order button.