TEST #3476 RESULTS
05/08/2026 - 6:07 PM
ACCELERATOR
MODEL
49.9
tokens/s
generation
1.02
sec
time to first token
1263
tokens/s
prompt
395
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 5 3600 6-Core Processor (znver2)
RAM
31.9GB
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
1514
tokens/s
55.6
tokens/s
694
ms
pp4096+tg256
1253
tokens/s
42.6
tokens/s
3.29
sec
pp2048+tg256
1248
tokens/s
47.6
tokens/s
1.66
sec
pp2048+tg768
1423
tokens/s
46.2
tokens/s
1.46
sec
pp1024+tg1024
1419
tokens/s
49.1
tokens/s
743
ms
pp1280+tg3072
1395
tokens/s
46.7
tokens/s
938
ms
pp384+tg1152
1526
tokens/s
53.4
tokens/s
270
ms
pp64+tg1024
1174
tokens/s
53.9
tokens/s
73
ms
pp16+tg1536
414
tokens/s
53.9
tokens/s
57
ms