Machine

Apple M1 Max · Apple M1 Max #55672f1c

63 runs of 18 models since Nov 2, 2026. Last run Nov 4, 2026.

ShareRedditX
63
runs
18
models
55
configs
8
configs run 2+ times

Hardware

GPU
Apple M1 Max 64 GB
CPU
Apple M1 Max 10 cores
RAM
64 GB
OS
macOS 27.2
Tools
LM Studio 63
Client
0.4.98
All Apple M1 Max benchmarks

Models on this machine

One dot per model, at its best config. Up is better answers, right is faster. Time to finish counts every token the model wrote, thinking included, so a model that talks a lot ends up further left.

Best trade-off: quality 60+, and nothing else here is both faster and betterOther models · dot size = number of runs
below 20 tok/s: slower than you read204060801002050100tok/s (log scale) →quality →outliers sit at the edges

Medians of each model's best config, 18 models.

Best config per model

sort

Runs with the same model, quant, tool, context, thinking and KV cache count as one config. Numbers are the median of its runs, and each dot is one run on the 0–100 quality scale. ★ marks the best config with at least 2 rated runs.

qwen/qwen3.6-35b-a3b
11 runs · 8 configs
★ best6bitLM Studio8k ctxthinking default
×2 runs
82.079.9–84.1
steady
57.7 tok/s
47.5–67.9
google/gemma-4-12b-qat
1 run · 1 config
★ bestQ4_0LM Studio32k ctxthinking on
74.7
1 rated run
27.8 tok/s
google/gemma-4-26b-a4b
1 run · 1 config
★ best4bitLM Studio8k ctxthinking default
73.7
1 rated run
48.6 tok/s
qwen_qwen3.5-4b@q8_0
2 runs · 2 configs
★ bestQ8_0LM Studio64k ctxthinking default
73.0
1 rated run
69.0 tok/s
qwen3.5-4b@q4_k_m
2 runs · 2 configs
★ bestQ4_K_MLM Studio64k ctxthinking default
72.9
1 rated run
49.1 tok/s
qwen3.5-4b@q6_k
2 runs · 2 configs
★ bestQ6_KLM Studio64k ctxthinking default
72.3
1 rated run
45.0 tok/s
qwen3.5-4b@q8_0
1 run · 1 config
★ bestQ8_0LM Studio64k ctxthinking default
69.9
1 rated run
45.4 tok/s
google/gemma-4-e2b
12 runs · 12 configs
★ best8bitLM Studio64k ctxthinking on
69.4
1 rated run
56.8 tok/s
qwen/qwen3.5-9b
2 runs · 2 configs
★ best4bitLM Studio64k ctxthinking default
67.1
1 rated run
48.5 tok/s
google/gemma-4-e4b
6 runs · 4 configs
★ best4bitLM Studio64k ctxthinking default
×2 runs
65.962.9–69.0
steady
46.6 tok/s
45.3–48.0

Compare with another machine

Run history