TEST #715 RESULTS
04/14/2025 - 7:26 AM
ACCELERATOR
MODEL
36.8
tokens/s
generation
995
ms
time to first token
1370
tokens/s
prompt
370
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 5900X 12-Core Processor (znver3)
RAM
47GB
OS
Linux
Kernel Release
6.13.10-200.fc41.x86_64
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #1 SMP PREEMPT_DYNAMIC Mon Apr 7 19:01:38 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1590
tokens/s
38.0
tokens/s
670
ms
pp4096+tg256
1207
tokens/s
33.6
tokens/s
3.42
sec
pp2048+tg256
1428
tokens/s
36.4
tokens/s
1.46
sec
pp2048+tg768
1428
tokens/s
36.0
tokens/s
1.46
sec
pp1024+tg1024
1559
tokens/s
37.2
tokens/s
683
ms
pp1280+tg3072
1510
tokens/s
35.2
tokens/s
874
ms
pp384+tg1152
1691
tokens/s
38.0
tokens/s
253
ms
pp64+tg1024
1444
tokens/s
38.7
tokens/s
69
ms
pp16+tg1536
473
tokens/s
38.3
tokens/s
59
ms