TEST #3558 RESULTS
05/17/2026 - 8:31 AM
ACCELERATOR
MODEL
34.7
tokens/s
generation
231
ms
time to first token
5678
tokens/s
prompt
949
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen AI 9 HX 370 w/ Radeon 890M
RAM
31.1GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
7414
tokens/s
33.4
tokens/s
167
ms
pp4096+tg256
6584
tokens/s
30.0
tokens/s
657
ms
pp2048+tg256
7366
tokens/s
36.6
tokens/s
308
ms
pp2048+tg768
7166
tokens/s
32.2
tokens/s
316
ms
pp1024+tg1024
7748
tokens/s
37.3
tokens/s
162
ms
pp1280+tg3072
7035
tokens/s
32.7
tokens/s
209
ms
pp384+tg1152
6259
tokens/s
37.2
tokens/s
99
ms
pp64+tg1024
1186
tokens/s
36.4
tokens/s
84
ms
pp16+tg1536
349
tokens/s
36.7
tokens/s
74
ms