TEST #523 RESULTS
04/07/2025 - 8:37 PM
ACCELERATOR
MODEL
66.1
tokens/s
generation
107
ms
time to first token
12782
tokens/s
prompt
1990
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 9800X3D 8-Core Processor
RAM
47.6GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
18967
tokens/s
69.0
tokens/s
68
ms
pp4096+tg256
14169
tokens/s
61.7
tokens/s
304
ms
pp2048+tg256
15499
tokens/s
64.1
tokens/s
148
ms
pp2048+tg768
14792
tokens/s
63.8
tokens/s
162
ms
pp1024+tg1024
16819
tokens/s
66.7
tokens/s
78
ms
pp1280+tg3072
15678
tokens/s
64.1
tokens/s
97
ms
pp384+tg1152
15419
tokens/s
68.1
tokens/s
40
ms
pp64+tg1024
2947
tokens/s
68.9
tokens/s
35
ms
pp16+tg1536
745
tokens/s
68.5
tokens/s
34
ms