TEST #4234 RESULTS
10/01/2026 - 2:55 AM
ACCELERATOR
MODEL
39.1
tokens/s
generation
832
ms
time to first token
1626
tokens/s
prompt
424
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i9-14900 (alderlake)
RAM
31.8GB
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
1915
tokens/s
39.9
tokens/s
560
ms
pp4096+tg256
1441
tokens/s
36.0
tokens/s
2.87
sec
pp2048+tg256
1718
tokens/s
39.0
tokens/s
1.22
sec
pp2048+tg768
1717
tokens/s
38.8
tokens/s
1.22
sec
pp1024+tg1024
1870
tokens/s
40.0
tokens/s
572
ms
pp1280+tg3072
1847
tokens/s
37.7
tokens/s
718
ms
pp384+tg1152
2106
tokens/s
40.3
tokens/s
207
ms
pp64+tg1024
1561
tokens/s
40.5
tokens/s
65
ms
pp16+tg1536
453
tokens/s
39.6
tokens/s
62
ms