TEST #3328 RESULTS
04/16/2026 - 6:40 AM
ACCELERATOR
MODEL
172
tokens/s
generation
90
ms
time to first token
16250
tokens/s
prompt
3143
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen1.5-0.5B-Chat-AWQ-fp16 - Q4_K - Medium
SYSTEM
CPU
RAM
7.8GB
OS
Linux
Kernel Release
6.18.5-nvidia-gpu
Architecture
aarch64
Version
Cosmopolitan 3.9.7 MODE=aarch64; #1 SMP Wed Jan 28 20:02:30 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
19396
tokens/s
192
tokens/s
58
ms
pp4096+tg256
11961
tokens/s
135
tokens/s
350
ms
pp2048+tg256
18029
tokens/s
159
tokens/s
120
ms
pp2048+tg768
18191
tokens/s
150
tokens/s
119
ms
pp1024+tg1024
23408
tokens/s
184
tokens/s
49
ms
pp1280+tg3072
22080
tokens/s
143
tokens/s
63
ms
pp384+tg1152
24190
tokens/s
194
tokens/s
21
ms
pp64+tg1024
7060
tokens/s
193
tokens/s
14
ms
pp16+tg1536
1936
tokens/s
194
tokens/s
13
ms