TEST #2368 RESULTS
01/27/2026 - 8:04 PM
ACCELERATOR
MODEL
128
tokens/s
generation
71
ms
time to first token
18389
tokens/s
prompt
3218
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 9950X3D 16-Core Processor
RAM
62.3GB
OS
Linux
Kernel Release
6.14.0-37-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #37~24.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Nov 20 10:25:38 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
25127
tokens/s
138
tokens/s
48
ms
pp4096+tg256
20522
tokens/s
105
tokens/s
210
ms
pp2048+tg256
22875
tokens/s
129
tokens/s
97
ms
pp2048+tg768
22790
tokens/s
128
tokens/s
98
ms
pp1024+tg1024
25119
tokens/s
132
tokens/s
48
ms
pp1280+tg3072
22468
tokens/s
121
tokens/s
64
ms
pp384+tg1152
21437
tokens/s
132
tokens/s
25
ms
pp64+tg1024
3941
tokens/s
133
tokens/s
23
ms
pp16+tg1536
1223
tokens/s
133
tokens/s
21
ms