TEST #3484 RESULTS
05/10/2026 - 10:44 AM
ACCELERATOR
MODEL
22.6
tokens/s
generation
490
ms
time to first token
2594
tokens/s
prompt
493
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 5 7500F 6-Core Processor (znver4)
RAM
31.6GB
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
3341
tokens/s
22.5
tokens/s
349
ms
pp4096+tg256
2838
tokens/s
20.6
tokens/s
1.49
sec
pp2048+tg256
3217
tokens/s
22.7
tokens/s
681
ms
pp2048+tg768
3221
tokens/s
23.3
tokens/s
681
ms
pp1024+tg1024
3406
tokens/s
24.4
tokens/s
341
ms
pp1280+tg3072
3134
tokens/s
22.7
tokens/s
450
ms
pp384+tg1152
3148
tokens/s
22.3
tokens/s
174
ms
pp64+tg1024
832
tokens/s
22.1
tokens/s
122
ms
pp16+tg1536
212
tokens/s
22.9
tokens/s
121
ms