TEST #1304 RESULTS
08/03/2025 - 6:12 PM
ACCELERATOR
MODEL
39.7
tokens/s
generation
1.06
sec
time to first token
1250
tokens/s
prompt
361
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 5800HS with Radeon Graphics (znver3)
RAM
15.4GB
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
1444
tokens/s
41.3
tokens/s
734
ms
pp4096+tg256
1160
tokens/s
36.0
tokens/s
3.56
sec
pp2048+tg256
1333
tokens/s
39.6
tokens/s
1.56
sec
pp2048+tg768
1329
tokens/s
39.1
tokens/s
1.57
sec
pp1024+tg1024
1444
tokens/s
40.3
tokens/s
733
ms
pp1280+tg3072
1394
tokens/s
38.1
tokens/s
942
ms
pp384+tg1152
1552
tokens/s
40.7
tokens/s
271
ms
pp64+tg1024
1167
tokens/s
41.4
tokens/s
78
ms
pp16+tg1536
430
tokens/s
41.2
tokens/s
59
ms