TEST #1425 RESULTS
08/25/2025 - 9:36 PM
ACCELERATOR
MODEL
85.1
tokens/s
generation
271
ms
time to first token
5172
tokens/s
prompt
1176
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen3 Mini Moe - Q4_K - Medium
SYSTEM
CPU
Intel Xeon W-3245 CPU @ 3.20GHz (cascadelake)
RAM
255.9GB
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
6662
tokens/s
86.6
tokens/s
166
ms
pp4096+tg256
4218
tokens/s
81.5
tokens/s
984
ms
pp2048+tg256
5555
tokens/s
84.8
tokens/s
381
ms
pp2048+tg768
5546
tokens/s
83.4
tokens/s
382
ms
pp1024+tg1024
6564
tokens/s
86.1
tokens/s
168
ms
pp1280+tg3072
6066
tokens/s
82.7
tokens/s
223
ms
pp384+tg1152
7630
tokens/s
86.2
tokens/s
63
ms
pp64+tg1024
3785
tokens/s
88.0
tokens/s
28
ms
pp16+tg1536
522
tokens/s
86.8
tokens/s
43
ms