TEST #3236 RESULTS
04/05/2026 - 8:40 AM
ACCELERATOR
MODEL
5.6
tokens/s
generation
7.06
sec
time to first token
186
tokens/s
prompt
53
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 8700G w/ Radeon 780M Graphics (znver4)
RAM
31.5GB
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
210
tokens/s
6.3
tokens/s
5.04
sec
pp4096+tg256
183
tokens/s
1.2
tokens/s
23.24
sec
pp2048+tg256
200
tokens/s
4.4
tokens/s
10.49
sec
pp2048+tg768
200
tokens/s
3.9
tokens/s
10.49
sec
pp1024+tg1024
210
tokens/s
5.6
tokens/s
5.03
sec
pp1280+tg3072
203
tokens/s
2.8
tokens/s
6.48
sec
pp384+tg1152
214
tokens/s
7.7
tokens/s
1.89
sec
pp64+tg1024
93
tokens/s
10.2
tokens/s
756
ms
pp16+tg1536
161
tokens/s
8.6
tokens/s
158
ms