Apple M6 and M5 Ultra joins the family - Welcome Apple's first ever 2nm chip

Table of Content

Apple M6 – Apple’s First Ever 2nm Chip
Apple M5 Ultra – Enter Quad-die with UltraFusion

After less than a year since the announcement of the Apple M5 series chips, today, Apple introduced their new additions to the family, the Apple M6 and Apple M5 Ultra that brings some powerful features to the Mac series line-up. This marks a significant advancement in performance and AI processing for Mac devices. The M6 is the company’s first 2 nanometer chip, designed for the new Mac mini. The M5 Ultra, built using a quad-die architecture, powers the new Mac Studio. Both chips are intended for users requiring high computational throughput and on-device AI capabilities.

Apple M6 – Apple’s First Ever 2nm Chip

Landing first into the Mac mini today is Apple’s first ever 2nm chip, the Apple M6. The chip features a 12-core CPU with two super cores, four performance cores, and six efficiency cores. It delivers enhanced single-threaded and multithreaded performance compared to the M5, with up to 1.2x faster multithreaded throughput and improved power efficiency.

The M6 includes a 12-core GPU with a Neural Accelerator in each core. This configuration enables up to 30 percent higher peak AI compute performance than the M5. The chip supports up to 32GB of unified memory and provides up to 170GB/s of memory bandwidth which is 10 percent more than the M5 and 2.5 times that of the M1.

The M6 also incorporates a Dual 16-core Neural Engine, offering up to 2x peak compute over prior generations. This enables faster execution of on-device AI models, including large language models used in agentic workflows. System frameworks automatically distribute workloads across the two neural engines for improved performance.

Apple M5 Ultra – Enter Quad-die with UltraFusion

Apple has also launched the M5 Ultra chip, the company’s most powerful processor to date, for use in the new Mac Studio. The chip uses a quad-die architecture built via UltraFusion technology, combining two dual-die M5 Max chips into a single system-on-a-chip. This design increases inter-die bandwidth to over 4.4TB/s and improves connection density by more than 6x.

The M5 Ultra offers up to 36-core CPU built with 12 super cores and 24 performance cores and up to a whopping 80-core GPU configurations. It provides up to 1.25x higher single-threaded performance and up to 1.3x higher multithreaded performance compared to the M3 Ultra. The GPU includes Neural Accelerators in every core, delivering up to 4.5x higher peak AI compute than the M3 Ultra.

The chip supports up to 512GB of unified memory and provides 1.2TB/s of memory bandwidth which is 50 percent more than the M3 Ultra. It includes a 32-core Neural Engine, advanced media engines for ProRes and AV1 decoding, and hardware-accelerated ray tracing. These features are optimized for complex workloads in 3D rendering, visual effects, scientific computing, and large-scale on-device AI inference.