TEST #3283 RESULTS
04/11/2026 - 6:24 AM
ACCELERATOR
MODEL
8.8
tokens/s
generation
59.69
sec
time to first token
23
tokens/s
prompt
15
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core Ultra 7 258V (arrowlake-s)
RAM
31.5GB
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
25
tokens/s
9.8
tokens/s
40.78
sec
pp4096+tg256
22
tokens/s
7.6
tokens/s
188.47
sec
pp2048+tg256
23
tokens/s
8.7
tokens/s
87.34
sec
pp2048+tg768
23
tokens/s
8.3
tokens/s
90.10
sec
pp1024+tg1024
23
tokens/s
8.2
tokens/s
43.70
sec
pp1280+tg3072
19
tokens/s
7.5
tokens/s
67.73
sec
pp384+tg1152
25
tokens/s
9.6
tokens/s
15.67
sec
pp64+tg1024
25
tokens/s
9.7
tokens/s
2.68
sec
pp16+tg1536
25
tokens/s
9.6
tokens/s
726
ms