TEST #4242 RESULTS
10/03/2026 - 3:00 PM
ACCELERATOR
MODEL
74.4
tokens/s
generation
756
ms
time to first token
2126
tokens/s
prompt
594
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i7 CPU 950 @ 3.07GHz (nehalem)
RAM
12GB
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
2450
tokens/s
88.0
tokens/s
430
ms
pp4096+tg256
1361
tokens/s
39.6
tokens/s
3.03
sec
pp2048+tg256
1926
tokens/s
63.4
tokens/s
1.08
sec
pp2048+tg768
1928
tokens/s
60.5
tokens/s
1.08
sec
pp1024+tg1024
2450
tokens/s
75.5
tokens/s
430
ms
pp1280+tg3072
2306
tokens/s
53.6
tokens/s
568
ms
pp384+tg1152
3214
tokens/s
89.1
tokens/s
130
ms
pp64+tg1024
2324
tokens/s
104
tokens/s
37
ms
pp16+tg1536
1171
tokens/s
96.4
tokens/s
23
ms