TEST #3391 RESULTS
04/26/2026 - 7:17 PM
ACCELERATOR
MODEL
6.2
tokens/s
generation
6.32
sec
time to first token
212
tokens/s
prompt
59
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 5 8500G w/ Radeon 740M Graphics (znver4)
RAM
31.2GB
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
213
tokens/s
7.4
tokens/s
4.95
sec
pp4096+tg256
192
tokens/s
1.5
tokens/s
22.03
sec
pp2048+tg256
232
tokens/s
5.2
tokens/s
9.04
sec
pp2048+tg768
231
tokens/s
4.7
tokens/s
9.08
sec
pp1024+tg1024
258
tokens/s
6.6
tokens/s
4.11
sec
pp1280+tg3072
242
tokens/s
3.3
tokens/s
5.44
sec
pp384+tg1152
284
tokens/s
8.1
tokens/s
1.44
sec
pp64+tg1024
129
tokens/s
10.2
tokens/s
569
ms
pp16+tg1536
126
tokens/s
9.0
tokens/s
192
ms