TEST #1636 RESULTS
10/11/2025 - 11:16 PM
ACCELERATOR
MODEL
53.1
tokens/s
generation
1.93
sec
time to first token
627
tokens/s
prompt
258
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M4 Max 12P+4E
RAM
128GB
OS
Darwin
Kernel Release
25.0.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.0.0: Wed Sep 17 21:41:26 PDT 2025; root:xnu-12377.1.9~141/RELEASE_ARM64_T6041
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
819
tokens/s
60.9
tokens/s
1.27
sec
pp4096+tg256
725
tokens/s
41.5
tokens/s
5.68
sec
pp2048+tg256
665
tokens/s
50.5
tokens/s
3.10
sec
pp2048+tg768
661
tokens/s
49.1
tokens/s
3.12
sec
pp1024+tg1024
696
tokens/s
53.9
tokens/s
1.49
sec
pp1280+tg3072
666
tokens/s
47.2
tokens/s
1.94
sec
pp384+tg1152
730
tokens/s
57.0
tokens/s
542
ms
pp64+tg1024
505
tokens/s
59.6
tokens/s
142
ms
pp16+tg1536
176
tokens/s
58.5
tokens/s
106
ms