TEST #2035 RESULTS
12/23/2025 - 6:26 PM
ACCELERATOR
MODEL
20.5
tokens/s
generation
526
ms
time to first token
2357
tokens/s
prompt
451
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon Gold 5418Y (sapphirerapids)
RAM
754.6GB
OS
Linux
Kernel Release
5.15.0-160-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #170-Ubuntu SMP Wed Oct 1 10:06:56 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
3035
tokens/s
21.4
tokens/s
384
ms
pp4096+tg256
2751
tokens/s
18.5
tokens/s
1.55
sec
pp2048+tg256
2914
tokens/s
19.9
tokens/s
753
ms
pp2048+tg768
2925
tokens/s
19.6
tokens/s
751
ms
pp1024+tg1024
3034
tokens/s
20.3
tokens/s
386
ms
pp1280+tg3072
2760
tokens/s
19.2
tokens/s
513
ms
pp384+tg1152
2739
tokens/s
21.0
tokens/s
180
ms
pp64+tg1024
810
tokens/s
22.5
tokens/s
118
ms
pp16+tg1536
240
tokens/s
22.1
tokens/s
105
ms