TEST #3438 RESULTS
05/04/2026 - 5:11 PM
ACCELERATOR
MODEL
59.8
tokens/s
generation
631
ms
time to first token
2030
tokens/s
prompt
578
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 5900X 12-Core Processor (znver3)
RAM
31.2GB
OS
Linux
Kernel Release
6.8.0-110-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #110-Ubuntu SMP PREEMPT_DYNAMIC Thu Mar 19 15:09:20 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2164
tokens/s
61.4
tokens/s
490
ms
pp4096+tg256
1989
tokens/s
53.8
tokens/s
2.08
sec
pp2048+tg256
2284
tokens/s
59.2
tokens/s
915
ms
pp2048+tg768
2283
tokens/s
58.8
tokens/s
916
ms
pp1024+tg1024
2448
tokens/s
60.8
tokens/s
434
ms
pp1280+tg3072
2375
tokens/s
57.1
tokens/s
555
ms
pp384+tg1152
2508
tokens/s
61.7
tokens/s
185
ms
pp64+tg1024
1765
tokens/s
63.1
tokens/s
51
ms
pp16+tg1536
456
tokens/s
62.5
tokens/s
50
ms