TEST #402 RESULTS
04/06/2025 - 6:09 PM
ACCELERATOR
MODEL
34.8
tokens/s
generation
1.19
sec
time to first token
1086
tokens/s
prompt
317
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen Threadripper 3960X 24-Core Processor (znver2)
RAM
251.5GB
OS
Linux
Kernel Release
6.8.0-57-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #59-Ubuntu SMP PREEMPT_DYNAMIC Sat Mar 15 17:40:59 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1299
tokens/s
37.4
tokens/s
815
ms
pp4096+tg256
1056
tokens/s
32.4
tokens/s
3.91
sec
pp2048+tg256
1194
tokens/s
35.2
tokens/s
1.74
sec
pp2048+tg768
1181
tokens/s
34.6
tokens/s
1.76
sec
pp1024+tg1024
1243
tokens/s
35.5
tokens/s
851
ms
pp1280+tg3072
1176
tokens/s
33.1
tokens/s
1.12
sec
pp384+tg1152
1262
tokens/s
35.3
tokens/s
331
ms
pp64+tg1024
1024
tokens/s
35.3
tokens/s
103
ms
pp16+tg1536
343
tokens/s
34.7
tokens/s
73
ms