TEST #1448 RESULTS
08/28/2025 - 3:54 PM
ACCELERATOR
MODEL
28.7
tokens/s
generation
2.05
sec
time to first token
758
tokens/s
prompt
220
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i7-7820X CPU @ 3.60GHz (skylake-avx512)
RAM
31.7GB
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
1029
tokens/s
34.2
tokens/s
1.03
sec
pp4096+tg256
436
tokens/s
25.6
tokens/s
9.43
sec
pp2048+tg256
931
tokens/s
31.2
tokens/s
2.23
sec
pp2048+tg768
920
tokens/s
30.3
tokens/s
2.26
sec
pp1024+tg1024
957
tokens/s
28.9
tokens/s
1.10
sec
pp1280+tg3072
774
tokens/s
24.1
tokens/s
1.69
sec
pp384+tg1152
875
tokens/s
27.5
tokens/s
475
ms
pp64+tg1024
654
tokens/s
28.4
tokens/s
132
ms
pp16+tg1536
249
tokens/s
28.3
tokens/s
95
ms