TEST #65 RESULTS
03/28/2025 - 8:56 PM
ACCELERATOR
MODEL
110
tokens/s
generation
336
ms
time to first token
3571
tokens/s
prompt
1054
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 7713 64-Core Processor (znver3)
RAM
1007.7GB
OS
Linux
Kernel Release
6.5.0-28-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #29~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Apr 4 14:39:20 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
4372
tokens/s
112
tokens/s
243
ms
pp4096+tg256
3906
tokens/s
102
tokens/s
1.06
sec
pp2048+tg256
4233
tokens/s
109
tokens/s
493
ms
pp2048+tg768
4213
tokens/s
107
tokens/s
495
ms
pp1024+tg1024
4363
tokens/s
111
tokens/s
244
ms
pp1280+tg3072
4304
tokens/s
106
tokens/s
307
ms
pp384+tg1152
3639
tokens/s
114
tokens/s
114
ms
pp64+tg1024
2467
tokens/s
116
tokens/s
34
ms
pp16+tg1536
638
tokens/s
115
tokens/s
33
ms