TEST #4135 RESULTS
09/09/2026 - 1:08 PM
ACCELERATOR
MODEL
60.4
tokens/s
generation
2.78
sec
time to first token
511
tokens/s
prompt
223
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 9800X3D 8-Core Processor
RAM
31.6GB
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
562
tokens/s
62.3
tokens/s
1.84
sec
pp4096+tg256
424
tokens/s
51.8
tokens/s
9.68
sec
pp2048+tg256
495
tokens/s
60.2
tokens/s
4.15
sec
pp2048+tg768
496
tokens/s
58.8
tokens/s
4.15
sec
pp1024+tg1024
537
tokens/s
60.9
tokens/s
1.92
sec
pp1280+tg3072
532
tokens/s
55.6
tokens/s
2.42
sec
pp384+tg1152
566
tokens/s
63.5
tokens/s
695
ms
pp64+tg1024
552
tokens/s
65.3
tokens/s
131
ms
pp16+tg1536
433
tokens/s
65.5
tokens/s
53
ms