TEST #3552 RESULTS
05/16/2026 - 9:01 AM
ACCELERATOR
MODEL
70.0
tokens/s
generation
443
ms
time to first token
2937
tokens/s
prompt
774
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i9-14900HX (alderlake)
RAM
15.7GB
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
3559
tokens/s
71.5
tokens/s
301
ms
pp4096+tg256
2757
tokens/s
66.7
tokens/s
1.50
sec
pp2048+tg256
3256
tokens/s
70.3
tokens/s
647
ms
pp2048+tg768
3243
tokens/s
68.7
tokens/s
647
ms
pp1024+tg1024
3477
tokens/s
69.4
tokens/s
309
ms
pp1280+tg3072
3408
tokens/s
66.6
tokens/s
389
ms
pp384+tg1152
3683
tokens/s
71.6
tokens/s
119
ms
pp64+tg1024
2254
tokens/s
72.8
tokens/s
43
ms
pp16+tg1536
794
tokens/s
72.7
tokens/s
33
ms