TEST #3698 RESULTS
06/08/2026 - 5:41 AM
ACCELERATOR
MODEL
24.2
tokens/s
generation
858
ms
time to first token
1824
tokens/s
prompt
372
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 9124 16-Core Processor (znver4)
RAM
62.6GB
OS
Linux
Kernel Release
6.8.0-71-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #71-Ubuntu SMP PREEMPT_DYNAMIC Tue Jul 22 16:52:38 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
992
tokens/s
24.8
tokens/s
1.07
sec
pp4096+tg256
1171
tokens/s
19.9
tokens/s
3.56
sec
pp2048+tg256
2398
tokens/s
23.7
tokens/s
896
ms
pp2048+tg768
2474
tokens/s
22.8
tokens/s
870
ms
pp1024+tg1024
2776
tokens/s
25.3
tokens/s
410
ms
pp1280+tg3072
2594
tokens/s
21.6
tokens/s
532
ms
pp384+tg1152
2803
tokens/s
26.1
tokens/s
174
ms
pp64+tg1024
974
tokens/s
27.0
tokens/s
98
ms
pp16+tg1536
237
tokens/s
26.6
tokens/s
104
ms