TEST #3628 RESULTS
05/26/2026 - 9:38 PM
ACCELERATOR
MODEL
68.1
tokens/s
generation
588
ms
time to first token
2187
tokens/s
prompt
633
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 7B13 64-Core Processor (znver3)
RAM
125.6GB
OS
Linux
Kernel Release
6.8.0-117-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #117~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu May 7 22:17:46 UTC
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2520
tokens/s
70.9
tokens/s
421
ms
pp4096+tg256
2162
tokens/s
60.5
tokens/s
1.91
sec
pp2048+tg256
2397
tokens/s
67.0
tokens/s
870
ms
pp2048+tg768
2386
tokens/s
65.9
tokens/s
874
ms
pp1024+tg1024
2591
tokens/s
69.0
tokens/s
409
ms
pp1280+tg3072
2515
tokens/s
64.6
tokens/s
523
ms
pp384+tg1152
2674
tokens/s
70.8
tokens/s
190
ms
pp64+tg1024
1917
tokens/s
72.4
tokens/s
47
ms
pp16+tg1536
524
tokens/s
71.8
tokens/s
44
ms