Apple has introduced two new computer chips promising to handle demanding work and artificial intelligence tasks directly on a Mac.
The M6 powers the new Mac mini, while the more powerful M5 Ultra is used in the new Mac Studio. Apple said both chips improve processing speed, graphics performance, and AI computing while keeping power use efficient.
The M6 is Apple’s first chip built using a 2-nanometer process. This allows Apple to fit more components into a smaller space, which can improve speed and energy efficiency.
It has a 12-core central processing unit, or CPU, which handles most computing tasks. This includes two “super cores” for jobs that need speed, four performance cores for heavier work, and six efficiency cores for routine and background tasks.
Apple said the M6 delivers up to 20% faster performance when using several CPU cores than the M5. It can also be up to 2.4 times faster than the M1.
The chip includes a 12-core graphics processor and two 16-core Neural Engines for AI work. Apple said its peak graphics computing power for AI is nearly 30% higher than the M5 and more than eight times that of the M1.
These improvements are meant to speed up tasks such as editing images, compiling code, searching newly added files, playing games, and running AI tools.
The M6 supports up to 32GB of unified memory, which is shared across the processor. It also offers memory bandwidth of up to 170GB per second, allowing the chip to move data more quickly.
For professionals with much heavier workloads, Apple introduced the M5 Ultra. The chip connects four processing dies so they work as one processor, a first for Apple silicon.
The M5 Ultra has up to a 36-core CPU, an 80-core graphics processor, and 512GB of unified memory. It provides memory bandwidth of up to 1.2TB per second, 50% higher than the M3 Ultra.
Apple said the additional power is intended for complex 3D graphics, visual effects, scientific research, high-resolution video editing, and large AI models. Its memory capacity allows some AI models with hundreds of billions of parameters to run directly on the Mac instead of relying on remote data centers.
Developers can use Apple tools, including Core AI, Core ML, Metal, and Xcode, to build, run, and fine-tune AI models locally on Macs equipped with the new chips.
