TEST #3504 RESULTS
05/11/2026 - 12:30 PM
ACCELERATOR
MODEL
70.9
tokens/s
generation
256
ms
time to first token
5472
tokens/s
prompt
1149
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i7-6700K CPU @ 4.00GHz (skylake)
RAM
39.9GB
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
7058
tokens/s
65.6
tokens/s
162
ms
pp4096+tg256
4445
tokens/s
69.5
tokens/s
939
ms
pp2048+tg256
6110
tokens/s
71.2
tokens/s
348
ms
pp2048+tg768
6110
tokens/s
72.2
tokens/s
351
ms
pp1024+tg1024
7130
tokens/s
72.1
tokens/s
157
ms
pp1280+tg3072
6779
tokens/s
72.5
tokens/s
202
ms
pp384+tg1152
7968
tokens/s
71.0
tokens/s
62
ms
pp64+tg1024
3207
tokens/s
71.5
tokens/s
33
ms
pp16+tg1536
440
tokens/s
72.0
tokens/s
49
ms