TEST #3081 RESULTS
03/14/2026 - 8:42 AM
ACCELERATOR
MODEL
18.2
tokens/s
generation
1.24
sec
time to first token
1106
tokens/s
prompt
253
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
RAM
119.7GB
OS
Linux
Kernel Release
6.11.0-1016-nvidia
Architecture
aarch64
Version
Cosmopolitan 3.9.7 MODE=aarch64; #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 16:52:46 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1374
tokens/s
19.1
tokens/s
800
ms
pp4096+tg256
948
tokens/s
16.6
tokens/s
4.39
sec
pp2048+tg256
1203
tokens/s
17.8
tokens/s
1.76
sec
pp2048+tg768
1195
tokens/s
17.8
tokens/s
1.77
sec
pp1024+tg1024
1347
tokens/s
18.4
tokens/s
820
ms
pp1280+tg3072
1309
tokens/s
17.5
tokens/s
1.03
sec
pp384+tg1152
1526
tokens/s
18.9
tokens/s
303
ms
pp64+tg1024
824
tokens/s
19.0
tokens/s
164
ms
pp16+tg1536
224
tokens/s
18.9
tokens/s
123
ms