TEST #1356 RESULTS
08/10/2025 - 8:54 PM
ACCELERATOR
MODEL
13.7
tokens/s
generation
9.07
sec
time to first token
137
tokens/s
prompt
59
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 4P+4E
RAM
16GB
OS
Darwin
Kernel Release
24.6.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.6.0: Mon Jul 14 11:29:54 PDT 2025; root:xnu-11417.140.69~1/RELEASE_ARM64_T8122
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
176
tokens/s
16.4
tokens/s
5.87
sec
pp4096+tg256
145
tokens/s
13.0
tokens/s
28.30
sec
pp2048+tg256
150
tokens/s
14.8
tokens/s
13.74
sec
pp2048+tg768
145
tokens/s
14.4
tokens/s
14.18
sec
pp1024+tg1024
147
tokens/s
14.7
tokens/s
7.03
sec
pp1280+tg3072
144
tokens/s
12.2
tokens/s
8.93
sec
pp384+tg1152
150
tokens/s
12.1
tokens/s
2.62
sec
pp64+tg1024
122
tokens/s
13.0
tokens/s
593
ms
pp16+tg1536
56
tokens/s
12.5
tokens/s
348
ms