Apple unveiled the M6 chip in a new Mac mini and the M5 Ultra in a new Mac Studio, its first 2-nanometer silicon and its first quad-die SoC, respectively. Apple lists the M6 with a 12-core CPU, a 12-core GPU carrying Neural Accelerators, a dual 16-core Neural Engine, and unified memory bandwidth that tops out at 170GB/s. M5 Ultra scales to a 36-core CPU and 80-core GPU with 1.2TB/s of bandwidth, which Apple says is 50 percent more than the M3 Ultra it replaces.
Apple’s own comparisons, not independently verified, are the only performance data released so far. Higher bandwidth and a larger Neural Engine matter specifically for running bigger language models locally rather than through a cloud API, a use case that has been building among developers on prior M-series Macs.
Apple Intelligence features tied to the new chips arrive with macOS 27 this fall, per the company, with no pricing or shipping date given for the hardware itself.
Apple’s newsroom announced the M6 and M5 Ultra chips on August 25, 2026.
