TEST #1081 RESULTS
06/20/2025 - 12:20 AM
ACCELERATOR
MODEL
75.7
tokens/s
generation
441
ms
time to first token
2792
tokens/s
prompt
783
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 5800X3D 8-Core Processor (znver3)
RAM
63.9GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
3009
tokens/s
80.8
tokens/s
353
ms
pp4096+tg256
2935
tokens/s
74.8
tokens/s
1.41
sec
pp2048+tg256
3277
tokens/s
68.4
tokens/s
638
ms
pp2048+tg768
3335
tokens/s
74.6
tokens/s
627
ms
pp1024+tg1024
3553
tokens/s
76.3
tokens/s
300
ms
pp1280+tg3072
3152
tokens/s
74.0
tokens/s
421
ms
pp384+tg1152
2926
tokens/s
74.7
tokens/s
143
ms
pp64+tg1024
2270
tokens/s
80.2
tokens/s
40
ms
pp16+tg1536
674
tokens/s
78.0
tokens/s
39
ms