Exploring Apple Silicon’s local AI performance with the Mac Studio and M4 Max — M4 Max beats GB10 and Strix Halo in decode throughput, but memory bandwidth isn't everything

Summary

Apple Silicon has been a popular choice for local AI exploration thanks to its high memory bandwidth compared to other unified memory platforms. We tested the M4 Max version of Apple's Mac Studio to see whether its 546GB/s of bandwidth makes it the clear winner in local LLM inference.