# Apple puts M6 and M5 Pro in the Mac mini

> Apple's smallest desktop gets a new M6 or M5 Pro, faster networking, and a much bigger pitch as an always-on local AI machine.

_Source: Apple official press release; performance figures are Apple-reported claims · 2026-08-25 · 5 min read · Verified against primary sources_

Canonical: https://iyu.app/e/mac-mini-m6-m5-pro

## The 60-second version

Apple's new Mac mini pairs M6 or M5 Pro silicon with a stronger pitch for local AI and pro workloads.

**Key points**

- M6 claims up to 4× faster AI performance than the previous M4 Mac mini, but the figure comes from Apple's own testing.
- M5 Pro scales to an 18-core CPU, 20-core GPU, and 64GB of unified memory for demanding workflows.
- Both models add Wi-Fi 7, Bluetooth 6, and faster Ethernet; M5 Pro adds Thunderbolt 5.
- US preorders start August 25 at $899 for M6 and $1,699 for M5 Pro; availability begins September 22.

**Verdict.** The hardware looks like a serious compact local-AI machine, but independent sustained-load tests and macOS 27 software support will determine whether the premium is justified.

## Full explainer

> **●** Apple announced the new Mac mini on August 25, 2026. Preorders begin today, and availability starts September 22.


### The short version — A tiny desktop with a bigger AI brief

Apple's new Mac mini comes in two tracks: an M6 model for everyday work and local AI, and an M5 Pro model for heavier creative, development, and research workloads. The physical idea is familiar; the performance target is much more ambitious.

- **4×** — Apple's claimed maximum M6 AI uplift vs. the previous M4 Mac mini
- **64GB** — maximum unified memory on the M5 Pro model
- **$899** — US starting price for M6 Mac mini


### What changed — Two chips, two buyer profiles

- **M6:** 12-core CPU and 12-core GPU; 16GB standard memory, configurable to 32GB; three Thunderbolt 4 ports.
- **M5 Pro:** Up to 18-core CPU and 20-core GPU; up to 64GB memory; three Thunderbolt 5 ports.
- **Both:** Wi-Fi 7, Bluetooth 6, 2.5Gb Ethernet, with 10Gb Ethernet available.
- **Availability:** Preorders start August 25; shipping begins September 22, 2026.
- **US pricing:** M6 from $899; M5 Pro from $1,699. Education pricing starts at $799 and $1,599.


### The AI angle — Local compute is the product story

The M6 adds Neural Accelerators to each GPU core and a dual 16-core Neural Engine. Apple says that combination makes AI workloads up to four times faster than on the prior M4 Mac mini. The M5 Pro is aimed at larger local models, image and video upscaling, and other workloads that benefit from more memory bandwidth.

> **⚑ Caveat:** The performance multipliers and speed comparisons are Apple's own measurements. They should be checked against independent tests using comparable memory, software versions, and workloads.


### Why it matters — A quiet box for always-on work

Apple is explicitly describing the Mac mini as an always-on desktop for agentic computing. That could mean local assistants, development tools, creative automation, or model-serving experiments that run beside a normal work setup. The hardware makes the idea more plausible; software support will decide how useful it becomes.


### What remains unknown — The launch announcement is not a review

Apple has not replaced independent testing. Buyers still need sustained-load results, thermals, noise measurements, actual SSD speeds, memory limits in AI workloads, and a clearer picture of which macOS 27 intelligence features run locally. The M5 Pro's higher ceiling may also make configuration choices more expensive, not merely faster.

> The new Mac mini is best understood as a local-compute platform in a small box — not just a faster office computer.


## Primary sources

- [Apple Newsroom](https://www.apple.com/newsroom/2026/08/apple-unveils-a-more-powerful-mac-mini-featuring-the-all-new-m6-and-m5-pro/)

---
_Published by iyu (https://iyu.app) — the day's AI news, checked against primary sources and rewritten in plain language. Free to quote with attribution and a link to the canonical URL._