TEST #2631 RESULTS
02/14/2026 - 2:55 PM
ACCELERATOR
MODEL
164
tokens/s
generation
65
ms
time to first token
22081
tokens/s
prompt
3824
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 3950X 16-Core Processor (znver2)
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
29808
tokens/s
160
tokens/s
41
ms
pp4096+tg256
18324
tokens/s
166
tokens/s
230
ms
pp2048+tg256
25258
tokens/s
164
tokens/s
87
ms
pp2048+tg768
25530
tokens/s
162
tokens/s
86
ms
pp1024+tg1024
31125
tokens/s
166
tokens/s
40
ms
pp1280+tg3072
29986
tokens/s
167
tokens/s
49
ms
pp384+tg1152
30564
tokens/s
164
tokens/s
19
ms
pp64+tg1024
6300
tokens/s
161
tokens/s
16
ms
pp16+tg1536
1832
tokens/s
164
tokens/s
15
ms