TEST #3400 RESULTS
04/27/2026 - 4:54 AM
ACCELERATOR
MODEL
21.5
tokens/s
generation
1.90
sec
time to first token
681
tokens/s
prompt
198
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core 5 210H (alderlake)
RAM
15.6GB
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
829
tokens/s
22.5
tokens/s
1.28
sec
pp4096+tg256
649
tokens/s
19.9
tokens/s
6.36
sec
pp2048+tg256
759
tokens/s
21.5
tokens/s
2.74
sec
pp2048+tg768
762
tokens/s
20.9
tokens/s
2.73
sec
pp1024+tg1024
764
tokens/s
21.2
tokens/s
1.38
sec
pp1280+tg3072
763
tokens/s
20.7
tokens/s
1.72
sec
pp384+tg1152
783
tokens/s
22.2
tokens/s
535
ms
pp64+tg1024
673
tokens/s
22.3
tokens/s
139
ms
pp16+tg1536
145
tokens/s
22.2
tokens/s
153
ms