Apple M3 4P+4E+10GPU
GPU
24
GB
PERFORMANCE OVERVIEW
Model
Llama 3.2 1B Instruct
Q4_K - Medium1.5B
Meta Llama 3.1 8B Instruct
Q4_K - Medium8.0B
Qwen2.5 14B Instruct
Q4_K - Medium14.8B
Prompt Speed
917tokens/s
110tokens/s
66tokens/s
Generation Speed
64.7tokens/s
10.2tokens/s
6.1tokens/s
Time to First Token
1.43sec
11.67sec
19.92sec
LocalScore
346
46
27
COMPARE MODELS
4 models tested
Select Models
Llama 3.2 1B Instruct
Q4_K - Medium
Meta Llama 3.1 8B Instruct
Q4_K - Medium
Qwen2.5 14B Instruct
Q4_K - Medium
Models Mlx Community Meta Llama 3.1 8B Instruct Bf16
Q6_K