M1 Pro (14-core GPU) — LLM Benchmarks

Measured LLM inference benchmarks for M1 Pro (14-core GPU) across all RAM configurations (16 GB, 32 GB). 6 benchmark rows across 3 models. Compare how RAM affects throughput. Real runs, not estimates.

6Benchmark rows

3Models tested

2RAM configurations

71.8Fastest 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 GBM1 Pro (14-core GPU, 16 GB)3 rows · 3 models71.8 tok/s peak 32 GBM1 Pro (14-core GPU, 32 GB)3 rows · 3 models71.0 tok/s peak

All benchmark rows — M1 Pro (14-core GPU)

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
M1 Pro (14-core GPU, 16 GB)	Llama 3.2 1B Instruct	Q4_K - Medium	—	71.8 tok/s	1040.1 tok/s	—	ref
M1 Pro (14-core GPU, 32 GB)	Llama 3.2 1B Instruct	Q4_K - Medium	—	71.0 tok/s	1063.8 tok/s	—	ref
M1 Pro (14-core GPU, 16 GB)	Llama 3.1 8B Instruct	Q4_K - Medium	—	20.1 tok/s	177.0 tok/s	—	ref
M1 Pro (14-core GPU, 32 GB)	Llama 3.1 8B Instruct	Q4_K - Medium	—	20.0 tok/s	173.1 tok/s	—	ref
M1 Pro (14-core GPU, 16 GB)	Qwen 2.5 14B Instruct	Q4_K - Medium	—	10.8 tok/s	92.5 tok/s	—	ref
M1 Pro (14-core GPU, 32 GB)	Qwen 2.5 14B Instruct	Q4_K - Medium	—	10.4 tok/s	88.9 tok/s	—	ref

Models tested on M1 Pro (14-core GPU)

Llama 3.1 8B Instruct Llama 3.2 1B Instruct Qwen 2.5 14B 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 →