TEST #616 RESULTS
04/08/2025 - 9:06 PM
ACCELERATOR
MODEL
101
tokens/s
generation
392
ms
time to first token
3829
tokens/s
prompt
997
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen Threadripper 2970WX 24-Core Processor
RAM
31.9GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
4855
tokens/s
116
tokens/s
220
ms
pp4096+tg256
2670
tokens/s
72.0
tokens/s
1.55
sec
pp2048+tg256
3733
tokens/s
85.1
tokens/s
560
ms
pp2048+tg768
3770
tokens/s
80.4
tokens/s
555
ms
pp1024+tg1024
4788
tokens/s
99.1
tokens/s
223
ms
pp1280+tg3072
4543
tokens/s
72.7
tokens/s
291
ms
pp384+tg1152
5530
tokens/s
115
tokens/s
80
ms
pp64+tg1024
2978
tokens/s
142
tokens/s
33
ms
pp16+tg1536
1593
tokens/s
131
tokens/s
19
ms