TEST #2415 RESULTS
01/31/2026 - 6:45 AM
ACCELERATOR
MODEL
78.2
tokens/s
generation
468
ms
time to first token
2911
tokens/s
prompt
786
LocalScore
HOW YOU STACK UP
Explore All ResultsSuperthoughts Lite v2 MOE Llama3.2 - Q4_K - Medium
SYSTEM
CPU
Intel Core i5-14450HX (alderlake)
RAM
15.7GB
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
3486
tokens/s
78.8
tokens/s
306
ms
pp4096+tg256
2492
tokens/s
74.8
tokens/s
1.66
sec
pp2048+tg256
3059
tokens/s
80.6
tokens/s
682
ms
pp2048+tg768
3060
tokens/s
78.6
tokens/s
682
ms
pp1024+tg1024
3473
tokens/s
80.3
tokens/s
308
ms
pp1280+tg3072
3305
tokens/s
77.0
tokens/s
400
ms
pp384+tg1152
4349
tokens/s
80.2
tokens/s
100
ms
pp64+tg1024
2335
tokens/s
77.1
tokens/s
39
ms
pp16+tg1536
642
tokens/s
76.0
tokens/s
36
ms