TEST #3993 RESULTS
08/05/2026 - 4:19 AM
MODEL
7.0
tokens/s
generation
74.29
sec
time to first token
19
tokens/s
prompt
12
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
RAM
35.3GB
OS
Linux
Kernel Release
6.18.15
Architecture
aarch64
Version
Cosmopolitan 3.9.7 MODE=aarch64; #1 SMP Tue Mar 17 01:36:53 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
20
tokens/s
7.9
tokens/s
51.60
sec
pp4096+tg256
17
tokens/s
6.1
tokens/s
237.68
sec
pp2048+tg256
18
tokens/s
6.7
tokens/s
115.92
sec
pp2048+tg768
18
tokens/s
6.5
tokens/s
116.34
sec
pp1024+tg1024
19
tokens/s
7.2
tokens/s
53.74
sec
pp1280+tg3072
19
tokens/s
6.6
tokens/s
68.52
sec
pp384+tg1152
19
tokens/s
7.3
tokens/s
20.38
sec
pp64+tg1024
19
tokens/s
7.2
tokens/s
3.50
sec
pp16+tg1536
20
tokens/s
7.4
tokens/s
934
ms