TEST #4235 RESULTS
10/01/2026 - 8:53 AM
ACCELERATOR
MODEL
18.5
tokens/s
generation
1.30
sec
time to first token
1089
tokens/s
prompt
250
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
RAM
121.7GB
OS
Linux
Kernel Release
7.0.0-1019-nvidia
Architecture
aarch64
Version
Cosmopolitan 3.9.7 MODE=aarch64; #19~24.04.2-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 6 18:35:11 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1267
tokens/s
18.2
tokens/s
871
ms
pp4096+tg256
878
tokens/s
16.2
tokens/s
4.72
sec
pp2048+tg256
1105
tokens/s
17.2
tokens/s
1.92
sec
pp2048+tg768
1170
tokens/s
19.1
tokens/s
1.80
sec
pp1024+tg1024
1402
tokens/s
19.8
tokens/s
780
ms
pp1280+tg3072
1355
tokens/s
17.4
tokens/s
996
ms
pp384+tg1152
1526
tokens/s
20.1
tokens/s
301
ms
pp64+tg1024
873
tokens/s
20.3
tokens/s
159
ms
pp16+tg1536
221
tokens/s
18.3
tokens/s
122
ms