TEST #3890 RESULTS
07/11/2026 - 1:30 PM
ACCELERATOR
MODEL
37.8
tokens/s
generation
5.75
sec
time to first token
260
tokens/s
prompt
120
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 7735HS with Radeon Graphics (znver3)
RAM
19.8GB
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
242
tokens/s
40.5
tokens/s
4.26
sec
pp4096+tg256
207
tokens/s
30.5
tokens/s
19.82
sec
pp2048+tg256
236
tokens/s
36.9
tokens/s
8.71
sec
pp2048+tg768
246
tokens/s
34.7
tokens/s
8.34
sec
pp1024+tg1024
249
tokens/s
38.9
tokens/s
4.14
sec
pp1280+tg3072
268
tokens/s
34.9
tokens/s
4.79
sec
pp384+tg1152
297
tokens/s
40.9
tokens/s
1.32
sec
pp64+tg1024
279
tokens/s
42.2
tokens/s
253
ms
pp16+tg1536
318
tokens/s
41.2
tokens/s
73
ms