M3 (GPU count not published) — LLM Benchmarks

Measured LLM inference benchmarks for M3 (GPU count not published) across all RAM configurations (16 GB). 2 benchmark rows across 2 models. Compare how RAM affects throughput. Real runs, not estimates.

2Benchmark rows

2Models tested

1RAM configurations

61.6Fastest avg tok/s

RAM configurations

Each configuration differs only in unified memory. More RAM = larger models fit. Throughput is similar across RAM tiers at the same model size.

16 GBM3 (GPU count not published, 16 GB)2 rows · 2 models61.6 tok/s peak

All benchmark rows — M3 (GPU count not published)

Sorted by avg tok/s descending. Click source badge to see original measurement.

Chip (RAM)	Model	Quant	RAM req.	Avg tok/s	Prompt tok/s	Runtime	Source
M3 (GPU count not published, 16 GB)	Llama 3.2 1B Instruct	Q4_K - Medium	—	61.6 tok/s	229.7 tok/s	—	ref
M3 (GPU count not published, 16 GB)	Llama 3.1 8B Instruct	Q4_K - Medium	—	11.8 tok/s	33.6 tok/s	—	ref

Models tested on M3 (GPU count not published)

Llama 3.1 8B Instruct Llama 3.2 1B Instruct

Data

benchmarks.json — full dataset · chips.json — chip summaries · benchmarks.csv — CSV export

Data sourced from factory lab measurements and community reference runs. See all chips →