Apple has officially announced its latest M-series chips, the Apple M6 and Apple M5 Ultra, targeting professional users who demand top-tier performance for AI modeling, 3D rendering, and complex video editing.

The Apple M6 is built on a cutting-edge 2nm process, allowing more transistors on a smaller die while improving power efficiency. It features a 12-core CPU (2 more than the M5) with 2 super cores, 4 performance cores, and 6 efficiency cores. Apple claims the M6 delivers the world's fastest single-threaded performance and is up to 1.2x faster in multi-thread tasks than the M5, and 2.4x faster than the M1.

On the GPU side, the M6 includes 12 cores with a Neural Accelerator, providing 30% faster peak GPU compute for AI tasks compared to the M5, and 8x faster than the M1. It also supports hardware-based ray tracing and Dynamic Caching. The M6's updated Neural Engine features a dual 16-core design, rated at 2x the peak compute of the M5 for on-device AI workflows. It supports up to 32GB of unified memory at 170GB/s speeds and debuts in the new Mac mini.

The Apple M5 Ultra takes a different approach by connecting two dual-die M5 Max chips via Apple's Ultra Fusion interconnect, creating a quad-die architecture. It is configurable with up to 36 CPU cores (12 super, 24 performance) and up to 80 GPU cores, plus 32 Neural Engine cores. Apple claims the M5 Ultra is 1.25x faster in single-thread and 1.3x faster in multi-thread performance than the M3 Ultra, with 4.3x the peak AI compute. It also features 2x more video encode/decode blocks, handling up to 33 streams of 8K ProRes video at 30fps, and supports up to 512GB of unified memory at 1.2TB/s speeds. The M5 Ultra debuts in the new Mac Studio.