TEST #557 RESULTS
04/08/2025 - 6:20 AM
ACCELERATOR
MODEL
82.5
tokens/s
generation
305
ms
time to first token
4285
tokens/s
prompt
1051
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 5900X 12-Core Processor (znver3)
RAM
127.9GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
5308
tokens/s
83.2
tokens/s
205
ms
pp4096+tg256
3975
tokens/s
77.5
tokens/s
1.04
sec
pp2048+tg256
4780
tokens/s
82.6
tokens/s
441
ms
pp2048+tg768
4760
tokens/s
80.3
tokens/s
442
ms
pp1024+tg1024
5124
tokens/s
83.8
tokens/s
212
ms
pp1280+tg3072
5226
tokens/s
81.8
tokens/s
257
ms
pp384+tg1152
5892
tokens/s
84.7
tokens/s
77
ms
pp64+tg1024
2565
tokens/s
84.9
tokens/s
37
ms
pp16+tg1536
933
tokens/s
84.0
tokens/s
29
ms