Apple M4 — Benchmarks & Specs

Bottom line: how fast is the Apple M4?

SpecPicks median: 20.0 tokens/sec across 3 cited single-user runs of 7-9B dense models at 4-bit (Q4) under ollama (range 16.9–21.2 tok/s), from Macyou (2 runs), LocalLLaMA. For local LLM inference it generates 11.7 tokens/sec running qwen2.5:14b at q4_K_M under ollama on a machine with 16 GB of unified memory, per Macyou. Unified memory is the binding constraint for local inference: the Apple M4 ships in configurations of up to 32 GB, per the manufacturer spec sheet, and a 14B model at q4_K_M needs roughly 8 GB for its weights alone (14B parameters at ~4.85 bits per weight, per llama.cpp), so it runs only on configurations with more unified memory than that. In PassMark CPU Mark it scores 22,203–23,643 points, across two PassMark results (PassMark CPU benchmark listing, PassMark CPU benchmark listing). Across all its cores it scores 13,845 points in Cinebench R23 multi, per Notebookcheck.

Every figure above is a row in the tables below, and each row links out to the review or public benchmark database the number was taken from; the memory ceiling and the weight-size estimate are not table rows and link to their sources inline; the SpecPicks median is computed here from the runs it names, each a row in the AI table. SpecPicks aggregates published measurements; it does not report first-party benchmark runs.