Chips & Hardware
Apple's new Mac Mini debuts with M6 chip, starts at $899
Apple unveiled a new Mac Mini running its M6 chip, with a $899 base model available for preorder now and shipping after September 22.
Apple on Tuesday unveiled a new Mac Mini powered by its latest processor, the M6. The base model comes with 256GB of storage and 16GB of RAM for $899, available for preorder now, though the device won’t ship until after September 22.
Apple says the new Mac Mini’s AI performance is four times better than the M4 model, with storage and graphics twice as fast. The M6 has a 12-core CPU and 12-core GPU, up from the 10-core CPU and 10-core GPU on the M4.
Alongside the M6 model, Apple is also selling a Mac Mini with the M5 Pro chip, which comes with 512GB of base storage and 24GB of RAM and starts at $1,699. The M5 Pro ships by default with a 15-core CPU and 16-core GPU, upgradable to an 18-core CPU and 20-core GPU. Both Mac Mini variants include Wi-Fi 7, Bluetooth 6, and a 2.5GB Ethernet port, with an option to upgrade to a 10GB port.
The new Mac Minis start shipping September 22 with macOS 27 and Siri AI. The line has grown popular over the past year among tech enthusiasts who use it to run local AI agents such as OpenClaw and Hermes, with some people reselling Mac Minis at marked-up prices on platforms like eBay earlier this year amid rising demand.
Apple also released new Mac Studio models running on the M5 Max and M5 Ultra chips, starting at $2,499 and $5,499, respectively. The M5 Max base model comes with 36GB of unified memory and 512GB of storage, while the M5 Ultra base model has 96GB of unified memory and 1TB of storage.
Earlier this year, Apple discontinued the previous base Mac Mini, which had 256GB of storage and retailed for $599, leaving the 512GB version with an M4 processor at $799 as the base model until now.
Why it matters
The refresh underscores how on-device AI performance has become Apple’s core selling point for the Mac line, as enthusiasts increasingly buy Mac Minis specifically to run local AI models rather than for general computing.