TEST #1600 RESULTS
10/07/2025 - 1:35 AM
ACCELERATOR
MODEL
38.1
tokens/s
generation
1.30
sec
time to first token
960
tokens/s
prompt
304
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 3700X 8-Core Processor (znver2)
RAM
62.8GB
OS
Linux
Kernel Release
6.6.87.2-microsoft-standard-WSL2
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #1 SMP PREEMPT_DYNAMIC Thu Jun 5 18:30:46 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1249
tokens/s
43.4
tokens/s
842
ms
pp4096+tg256
992
tokens/s
33.0
tokens/s
4.16
sec
pp2048+tg256
1103
tokens/s
36.9
tokens/s
1.88
sec
pp2048+tg768
1128
tokens/s
35.4
tokens/s
1.84
sec
pp1024+tg1024
1172
tokens/s
37.4
tokens/s
898
ms
pp1280+tg3072
913
tokens/s
33.0
tokens/s
1.43
sec
pp384+tg1152
843
tokens/s
38.6
tokens/s
479
ms
pp64+tg1024
943
tokens/s
41.7
tokens/s
91
ms
pp16+tg1536
298
tokens/s
43.4
tokens/s
76
ms