TEST #1918 RESULTS
12/08/2025 - 10:51 AM
ACCELERATOR
MODEL
62.1
tokens/s
generation
110
ms
time to first token
12666
tokens/s
prompt
1924
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon Gold 6426Y (sapphirerapids)
RAM
128GB
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
18704
tokens/s
60.8
tokens/s
71
ms
pp4096+tg256
13699
tokens/s
60.2
tokens/s
317
ms
pp2048+tg256
16705
tokens/s
62.5
tokens/s
140
ms
pp2048+tg768
16562
tokens/s
64.8
tokens/s
141
ms
pp1024+tg1024
17623
tokens/s
61.3
tokens/s
72
ms
pp1280+tg3072
16007
tokens/s
60.2
tokens/s
97
ms
pp384+tg1152
12917
tokens/s
61.3
tokens/s
45
ms
pp64+tg1024
1126
tokens/s
62.6
tokens/s
71
ms
pp16+tg1536
647
tokens/s
64.9
tokens/s
40
ms