TEST #3302 RESULTS
04/13/2026 - 6:44 PM
ACCELERATOR
MODEL
61.6
tokens/s
generation
619
ms
time to first token
2083
tokens/s
prompt
592
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
6.18.7-76061807-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #202601231045~1769703228~24.04~cb87b5b SMP PREEMPT_DYNAMIC Thu J
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2455
tokens/s
64.3
tokens/s
433
ms
pp4096+tg256
2023
tokens/s
55.2
tokens/s
2.04
sec
pp2048+tg256
2284
tokens/s
60.6
tokens/s
915
ms
pp2048+tg768
2280
tokens/s
59.9
tokens/s
916
ms
pp1024+tg1024
2441
tokens/s
62.5
tokens/s
435
ms
pp1280+tg3072
2375
tokens/s
58.6
tokens/s
555
ms
pp384+tg1152
2495
tokens/s
63.9
tokens/s
180
ms
pp64+tg1024
1848
tokens/s
65.0
tokens/s
49
ms
pp16+tg1536
544
tokens/s
64.3
tokens/s
44
ms