TEST #2632 RESULTS
02/14/2026 - 3:02 PM
ACCELERATOR
MODEL
80.5
tokens/s
generation
188
ms
time to first token
6998
tokens/s
prompt
1441
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 3950X 16-Core Processor (znver2)
RAM
63.9GB
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
8998
tokens/s
80.7
tokens/s
126
ms
pp4096+tg256
6545
tokens/s
80.8
tokens/s
639
ms
pp2048+tg256
8154
tokens/s
81.3
tokens/s
264
ms
pp2048+tg768
7986
tokens/s
81.7
tokens/s
269
ms
pp1024+tg1024
9152
tokens/s
79.9
tokens/s
125
ms
pp1280+tg3072
8951
tokens/s
79.9
tokens/s
155
ms
pp384+tg1152
8859
tokens/s
80.2
tokens/s
56
ms
pp64+tg1024
3382
tokens/s
81.3
tokens/s
31
ms
pp16+tg1536
955
tokens/s
78.8
tokens/s
29
ms