M1 Ultra (48-core GPU) — LLM Benchmarks

Measured LLM inference benchmarks for M1 Ultra (48-core GPU) across all RAM configurations (128 GB). 3 benchmark rows across 3 models. Compare how RAM affects throughput. Real runs, not estimates.

3Benchmark rows

3Models tested

1RAM configurations

138.0Fastest 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.

128 GBM1 Ultra (48-core GPU, 128 GB)3 rows · 3 models138.0 tok/s peak

All benchmark rows — M1 Ultra (48-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 Ultra (48-core GPU, 128 GB)	Llama 3.2 1B Instruct	Q4_K - Medium	—	138.0 tok/s	2582.4 tok/s	—	ref
M1 Ultra (48-core GPU, 128 GB)	Llama 3.1 8B Instruct	Q4_K - Medium	—	48.9 tok/s	534.4 tok/s	—	ref
M1 Ultra (48-core GPU, 128 GB)	Qwen 2.5 14B Instruct	Q4_K - Medium	—	27.8 tok/s	289.9 tok/s	—	ref

Models tested on M1 Ultra (48-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 →