TEST #1306 RESULTS
08/03/2025 - 10:09 PM
ACCELERATOR
MODEL
107
tokens/s
generation
100
ms
time to first token
13136
tokens/s
prompt
2412
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 9700X 8-Core Processor
RAM
123.4GB
OS
Linux
Kernel Release
6.8.0-71-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #71-Ubuntu SMP PREEMPT_DYNAMIC Tue Jul 22 16:52:38 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
17663
tokens/s
123
tokens/s
66
ms
pp4096+tg256
13312
tokens/s
77.0
tokens/s
321
ms
pp2048+tg256
15685
tokens/s
105
tokens/s
140
ms
pp2048+tg768
15620
tokens/s
97.3
tokens/s
141
ms
pp1024+tg1024
17649
tokens/s
107
tokens/s
66
ms
pp1280+tg3072
15826
tokens/s
92.6
tokens/s
90
ms
pp384+tg1152
17069
tokens/s
115
tokens/s
31
ms
pp64+tg1024
4234
tokens/s
124
tokens/s
22
ms
pp16+tg1536
1164
tokens/s
118
tokens/s
21
ms