TEST #916 RESULTS
05/19/2025 - 8:20 AM
ACCELERATOR
MODEL
5.4
tokens/s
generation
2.40
sec
time to first token
587
tokens/s
prompt
110
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 7900X 12-Core Processor (znver4)
RAM
62GB
OS
Linux
Kernel Release
6.5.0-41-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #41~22.04.2-Ubuntu SMP PREEMPT_DYNAMIC Mon Jun 3 11:32:55 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
734
tokens/s
6.1
tokens/s
1.56
sec
pp4096+tg256
525
tokens/s
1.5
tokens/s
8.43
sec
pp2048+tg256
638
tokens/s
4.0
tokens/s
3.49
sec
pp2048+tg768
640
tokens/s
3.6
tokens/s
3.48
sec
pp1024+tg1024
719
tokens/s
5.3
tokens/s
1.59
sec
pp1280+tg3072
675
tokens/s
2.9
tokens/s
2.09
sec
pp384+tg1152
744
tokens/s
7.3
tokens/s
616
ms
pp64+tg1024
445
tokens/s
9.7
tokens/s
210
ms
pp16+tg1536
161
tokens/s
8.2
tokens/s
154
ms