TEST #3256 RESULTS
04/08/2026 - 12:52 PM
ACCELERATOR
MODEL
29.5
tokens/s
generation
1.22
sec
time to first token
1151
tokens/s
prompt
303
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core Ultra 9 185H (alderlake)
RAM
31.4GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1221
tokens/s
29.5
tokens/s
872
ms
pp4096+tg256
970
tokens/s
28.0
tokens/s
4.26
sec
pp2048+tg256
1129
tokens/s
28.6
tokens/s
1.85
sec
pp2048+tg768
1134
tokens/s
28.2
tokens/s
1.84
sec
pp1024+tg1024
1350
tokens/s
29.0
tokens/s
791
ms
pp1280+tg3072
1348
tokens/s
31.1
tokens/s
982
ms
pp384+tg1152
1622
tokens/s
30.4
tokens/s
266
ms
pp64+tg1024
1217
tokens/s
30.8
tokens/s
85
ms
pp16+tg1536
366
tokens/s
30.0
tokens/s
72
ms