TEST #781 RESULTS
04/29/2025 - 1:42 PM
ACCELERATOR
MODEL
18.6
tokens/s
generation
8.64
sec
time to first token
180
tokens/s
prompt
73
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i7-10870H CPU @ 2.20GHz (skylake)
RAM
15.8GB
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
194
tokens/s
20.2
tokens/s
5.32
sec
pp4096+tg256
129
tokens/s
15.6
tokens/s
31.82
sec
pp2048+tg256
168
tokens/s
18.3
tokens/s
12.21
sec
pp2048+tg768
160
tokens/s
17.9
tokens/s
12.83
sec
pp1024+tg1024
195
tokens/s
18.6
tokens/s
5.30
sec
pp1280+tg3072
162
tokens/s
17.0
tokens/s
7.93
sec
pp384+tg1152
211
tokens/s
19.7
tokens/s
1.87
sec
pp64+tg1024
211
tokens/s
20.2
tokens/s
349
ms
pp16+tg1536
186
tokens/s
20.1
tokens/s
132
ms