TEST #3397 RESULTS
04/27/2026 - 3:32 AM
ACCELERATOR
MODEL
11.5
tokens/s
generation
854
ms
time to first token
1908
tokens/s
prompt
295
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 5 8600G w/ Radeon 760M Graphics (znver4)
RAM
31.1GB
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
2350
tokens/s
12.7
tokens/s
514
ms
pp4096+tg256
1375
tokens/s
4.0
tokens/s
3.22
sec
pp2048+tg256
1892
tokens/s
8.2
tokens/s
1.22
sec
pp2048+tg768
1886
tokens/s
7.5
tokens/s
1.22
sec
pp1024+tg1024
2361
tokens/s
10.8
tokens/s
514
ms
pp1280+tg3072
2140
tokens/s
6.5
tokens/s
692
ms
pp384+tg1152
2995
tokens/s
15.2
tokens/s
176
ms
pp64+tg1024
1683
tokens/s
21.1
tokens/s
67
ms
pp16+tg1536
491
tokens/s
17.5
tokens/s
58
ms