TEST #116 RESULTS
03/28/2025 - 9:54 PM
ACCELERATOR
MODEL
74.0
tokens/s
generation
601
ms
time to first token
2102
tokens/s
prompt
637
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
62.7GB
OS
Linux
Kernel Release
6.5.0-21-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #21~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Fri Feb 9 13:32:52 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2182
tokens/s
78.4
tokens/s
482
ms
pp4096+tg256
2075
tokens/s
61.7
tokens/s
1.99
sec
pp2048+tg256
2400
tokens/s
70.9
tokens/s
867
ms
pp2048+tg768
2402
tokens/s
69.8
tokens/s
867
ms
pp1024+tg1024
2569
tokens/s
74.9
tokens/s
411
ms
pp1280+tg3072
2493
tokens/s
67.3
tokens/s
527
ms
pp384+tg1152
2340
tokens/s
79.2
tokens/s
176
ms
pp64+tg1024
1872
tokens/s
82.4
tokens/s
46
ms
pp16+tg1536
583
tokens/s
80.9
tokens/s
39
ms