TEST #3109 RESULTS
03/18/2026 - 4:38 PM
ACCELERATOR
MODEL
39.5
tokens/s
generation
762
ms
time to first token
1707
tokens/s
prompt
446
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon Gold 6248R CPU @ 3.00GHz (cascadelake)
RAM
128GB
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
2049
tokens/s
39.7
tokens/s
525
ms
pp4096+tg256
1609
tokens/s
36.3
tokens/s
2.57
sec
pp2048+tg256
1908
tokens/s
39.4
tokens/s
1.10
sec
pp2048+tg768
1896
tokens/s
39.2
tokens/s
1.11
sec
pp1024+tg1024
2057
tokens/s
40.1
tokens/s
523
ms
pp1280+tg3072
1981
tokens/s
38.5
tokens/s
671
ms
pp384+tg1152
2080
tokens/s
40.3
tokens/s
227
ms
pp64+tg1024
1412
tokens/s
41.3
tokens/s
69
ms
pp16+tg1536
375
tokens/s
40.9
tokens/s
66
ms