M1 — benchmark record

M1 local LLM benchmarks across 8 GB, 16 GB RAM tiers on Apple Silicon Mac. 9 published rows across 9 models with explicit evidence state and RAM-tier comparison. Peak published speed is 58.0 tok/s.

9Benchmark rows

9Models tested

2RAM configurations

58.0Fastest 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.

8 GBM1 (8 GB)3 rows · 3 models58.0 tok/s peak 16 GBM1 (16 GB)6 rows · 6 models58.0 tok/s peak

All benchmark rows — M1

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 (8 GB)	Gemma 4 E2B	Q4_K - Medium	—	58.0 tok/s	—	Ollama	ref
M1 (16 GB)	Phi-4 Mini Instruct 3.8B	Q4_K - Medium	—	58.0 tok/s	—	Ollama	ref
M1 (8 GB)	Gemma 3 4B	Q4_K - Medium	—	48.0 tok/s	—	Ollama	ref
M1 (8 GB)	Gemma 4 E4B	Q4_K - Medium	—	42.0 tok/s	—	Ollama	ref
M1 (16 GB)	Mistral 7B v0.3	Q4_K - Medium	—	42.0 tok/s	—	Ollama	ref
M1 (16 GB)	Llama 3.1 8B	Q4_K - Medium	—	40.0 tok/s	—	Ollama	ref
M1 (16 GB)	Qwen 3 8B	Q4_K - Medium	—	38.0 tok/s	—	Ollama	ref
M1 (16 GB)	DeepSeek R1 Distill Llama 8B	Q4_K - Medium	—	38.0 tok/s	—	Ollama	ref
M1 (16 GB)	Qwen3.5-9B	Q4_K - Medium	—	35.0 tok/s	—	Ollama	ref

Models tested on M1

DeepSeek R1 Distill Llama 8B Gemma 3 4B Gemma 4 E2B Gemma 4 E4B Llama 3.1 8B Mistral 7B v0.3 Phi-4 Mini Instruct 3.8B Qwen 3 8B Qwen3.5-9B

Data

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

Data from in-house lab measurements plus community-published benchmarks. See all chip families →