TEST #4133 RESULTS
09/08/2026 - 8:03 AM
ACCELERATOR
MODEL
16.5
tokens/s
generation
35.26
sec
time to first token
40
tokens/s
prompt
26
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen Embedded V1500B (znver2)
RAM
31.3GB
OS
Linux
Kernel Release
4.4.302+
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #90080 SMP Wed Jul 22 00:10:11 CST 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
43
tokens/s
18.5
tokens/s
24.01
sec
pp4096+tg256
34
tokens/s
12.4
tokens/s
120.76
sec
pp2048+tg256
39
tokens/s
15.4
tokens/s
52.71
sec
pp2048+tg768
39
tokens/s
15.3
tokens/s
53.16
sec
pp1024+tg1024
41
tokens/s
16.9
tokens/s
24.83
sec
pp1280+tg3072
41
tokens/s
14.3
tokens/s
31.22
sec
pp384+tg1152
44
tokens/s
18.1
tokens/s
8.68
sec
pp64+tg1024
46
tokens/s
18.9
tokens/s
1.43
sec
pp16+tg1536
31
tokens/s
18.6
tokens/s
561
ms