TEST #1462 RESULTS
08/30/2025 - 10:09 AM
ACCELERATOR
MODEL
52.3
tokens/s
generation
699
ms
time to first token
1885
tokens/s
prompt
521
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 5900HX with Radeon Graphics (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
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2191
tokens/s
53.3
tokens/s
486
ms
pp4096+tg256
1768
tokens/s
49.3
tokens/s
2.34
sec
pp2048+tg256
2002
tokens/s
53.1
tokens/s
1.04
sec
pp2048+tg768
2025
tokens/s
52.3
tokens/s
1.03
sec
pp1024+tg1024
2178
tokens/s
52.3
tokens/s
489
ms
pp1280+tg3072
2123
tokens/s
51.3
tokens/s
621
ms
pp384+tg1152
2382
tokens/s
53.1
tokens/s
180
ms
pp64+tg1024
1744
tokens/s
53.1
tokens/s
55
ms
pp16+tg1536
552
tokens/s
52.9
tokens/s
47
ms