TEST #1965 RESULTS
12/18/2025 - 5:02 PM
ACCELERATOR
MODEL
3.3
tokens/s
generation
13.58
sec
time to first token
94
tokens/s
prompt
28
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 8845HS w/ Radeon 780M Graphics (znver4)
RAM
31.3GB
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
107
tokens/s
3.8
tokens/s
9.85
sec
pp4096+tg256
95
tokens/s
0.8
tokens/s
44.37
sec
pp2048+tg256
107
tokens/s
2.8
tokens/s
19.53
sec
pp2048+tg768
106
tokens/s
2.5
tokens/s
19.71
sec
pp1024+tg1024
108
tokens/s
3.4
tokens/s
9.74
sec
pp1280+tg3072
105
tokens/s
1.7
tokens/s
12.46
sec
pp384+tg1152
82
tokens/s
4.5
tokens/s
4.91
sec
pp64+tg1024
53
tokens/s
5.6
tokens/s
1.34
sec
pp16+tg1536
81
tokens/s
4.9
tokens/s
321
ms