TEST #3554 RESULTS
05/16/2026 - 3:35 PM
ACCELERATOR
MODEL
71.3
tokens/s
generation
438
ms
time to first token
3104
tokens/s
prompt
797
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core Ultra 9 275HX (arrowlake-s)
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
3802
tokens/s
59.7
tokens/s
285
ms
pp4096+tg256
2628
tokens/s
67.1
tokens/s
1.57
sec
pp2048+tg256
3390
tokens/s
73.2
tokens/s
618
ms
pp2048+tg768
3389
tokens/s
71.6
tokens/s
619
ms
pp1024+tg1024
3483
tokens/s
73.8
tokens/s
309
ms
pp1280+tg3072
3656
tokens/s
68.9
tokens/s
364
ms
pp384+tg1152
4174
tokens/s
75.2
tokens/s
106
ms
pp64+tg1024
2451
tokens/s
77.2
tokens/s
38
ms
pp16+tg1536
959
tokens/s
75.3
tokens/s
27
ms