TEST #3313 RESULTS
04/15/2026 - 10:18 AM
ACCELERATOR
MODEL
18.8
tokens/s
generation
1.33
sec
time to first token
963
tokens/s
prompt
239
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon w3-2435 (sapphirerapids)
RAM
31GB
OS
Linux
Kernel Release
6.17.0-1011-oem
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #11-Ubuntu SMP PREEMPT_DYNAMIC Fri Jan 23 06:18:43 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1220
tokens/s
21.4
tokens/s
889
ms
pp4096+tg256
956
tokens/s
15.1
tokens/s
4.35
sec
pp2048+tg256
1096
tokens/s
18.2
tokens/s
1.93
sec
pp2048+tg768
1080
tokens/s
17.4
tokens/s
1.95
sec
pp1024+tg1024
1160
tokens/s
19.0
tokens/s
936
ms
pp1280+tg3072
1090
tokens/s
16.5
tokens/s
1.23
sec
pp384+tg1152
1194
tokens/s
20.1
tokens/s
382
ms
pp64+tg1024
650
tokens/s
20.9
tokens/s
143
ms
pp16+tg1536
218
tokens/s
20.5
tokens/s
119
ms