TEST #2974 RESULTS
03/05/2026 - 10:41 AM
ACCELERATOR
MODEL
118
tokens/s
generation
326
ms
time to first token
4655
tokens/s
prompt
1190
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
63.9GB
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
5980
tokens/s
132
tokens/s
180
ms
pp4096+tg256
3155
tokens/s
88.7
tokens/s
1.31
sec
pp2048+tg256
4536
tokens/s
96.4
tokens/s
462
ms
pp2048+tg768
4531
tokens/s
89.3
tokens/s
463
ms
pp1024+tg1024
5866
tokens/s
116
tokens/s
182
ms
pp1280+tg3072
5602
tokens/s
82.9
tokens/s
237
ms
pp384+tg1152
6919
tokens/s
140
tokens/s
61
ms
pp64+tg1024
3496
tokens/s
165
tokens/s
23
ms
pp16+tg1536
1808
tokens/s
150
tokens/s
14
ms