TEST #409 RESULTS
04/06/2025 - 7:15 PM
ACCELERATOR
MODEL
54.4
tokens/s
generation
696
ms
time to first token
1861
tokens/s
prompt
526
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 5800X 8-Core Processor (znver3)
RAM
31.9GB
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
2214
tokens/s
54.9
tokens/s
481
ms
pp4096+tg256
1755
tokens/s
52.8
tokens/s
2.35
sec
pp2048+tg256
2060
tokens/s
54.8
tokens/s
1.01
sec
pp2048+tg768
2027
tokens/s
54.3
tokens/s
1.03
sec
pp1024+tg1024
2225
tokens/s
55.1
tokens/s
478
ms
pp1280+tg3072
2141
tokens/s
53.4
tokens/s
616
ms
pp384+tg1152
2216
tokens/s
54.7
tokens/s
192
ms
pp64+tg1024
1595
tokens/s
54.5
tokens/s
58
ms
pp16+tg1536
518
tokens/s
54.7
tokens/s
49
ms