TEST #1258 RESULTS
07/27/2025 - 4:22 PM
ACCELERATOR
MODEL
211
tokens/s
generation
229
ms
time to first token
6061
tokens/s
prompt
1774
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 5900X 12-Core Processor (znver3)
RAM
31.2GB
OS
Linux
Kernel Release
6.8.0-64-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #67-Ubuntu SMP PREEMPT_DYNAMIC Sun Jun 15 20:23:31 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
7479
tokens/s
226
tokens/s
141
ms
pp4096+tg256
4849
tokens/s
168
tokens/s
851
ms
pp2048+tg256
6351
tokens/s
201
tokens/s
328
ms
pp2048+tg768
6336
tokens/s
196
tokens/s
328
ms
pp1024+tg1024
7483
tokens/s
214
tokens/s
141
ms
pp1280+tg3072
7136
tokens/s
191
tokens/s
184
ms
pp384+tg1152
8455
tokens/s
228
tokens/s
50
ms
pp64+tg1024
5343
tokens/s
238
tokens/s
16
ms
pp16+tg1536
1114
tokens/s
233
tokens/s
21
ms