TEST #3793 RESULTS
06/26/2026 - 9:12 AM
ACCELERATOR
MODEL
7.6
tokens/s
generation
1.62
sec
time to first token
829
tokens/s
prompt
157
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 9700X 8-Core Processor
RAM
61.6GB
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
1069
tokens/s
7.7
tokens/s
1.09
sec
pp4096+tg256
819
tokens/s
2.1
tokens/s
5.46
sec
pp2048+tg256
949
tokens/s
5.5
tokens/s
2.37
sec
pp2048+tg768
948
tokens/s
5.0
tokens/s
2.37
sec
pp1024+tg1024
1042
tokens/s
7.3
tokens/s
1.11
sec
pp1280+tg3072
970
tokens/s
4.2
tokens/s
1.47
sec
pp384+tg1152
1046
tokens/s
10.4
tokens/s
434
ms
pp64+tg1024
339
tokens/s
14.6
tokens/s
230
ms
pp16+tg1536
277
tokens/s
11.9
tokens/s
93
ms