TEST #961 RESULTS
05/31/2025 - 7:34 PM
ACCELERATOR
MODEL
15.9
tokens/s
generation
8.26
sec
time to first token
146
tokens/s
prompt
65
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M4 4P+6E
RAM
32GB
OS
Darwin
Kernel Release
24.5.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.5.0: Tue Apr 22 19:54:43 PDT 2025; root:xnu-11417.121.6~2/RELEASE_ARM64_T8132
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
183
tokens/s
18.5
tokens/s
5.64
sec
pp4096+tg256
162
tokens/s
14.7
tokens/s
25.40
sec
pp2048+tg256
170
tokens/s
17.0
tokens/s
12.08
sec
pp2048+tg768
163
tokens/s
16.0
tokens/s
12.59
sec
pp1024+tg1024
162
tokens/s
16.2
tokens/s
6.39
sec
pp1280+tg3072
151
tokens/s
13.8
tokens/s
8.51
sec
pp384+tg1152
140
tokens/s
15.5
tokens/s
2.80
sec
pp64+tg1024
126
tokens/s
15.8
tokens/s
569
ms
pp16+tg1536
56
tokens/s
15.3
tokens/s
344
ms