TEST #71 RESULTS
03/28/2025 - 9:04 PM
ACCELERATOR
MODEL
120
tokens/s
generation
236
ms
time to first token
5202
tokens/s
prompt
1382
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon Platinum 8352Y CPU @ 2.20GHz (icelake-server)
RAM
2015.5GB
OS
Linux
Kernel Release
6.5.0-35-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #35~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Tue May 7 09:00:52 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
6254
tokens/s
109
tokens/s
172
ms
pp4096+tg256
5462
tokens/s
110
tokens/s
761
ms
pp2048+tg256
6082
tokens/s
120
tokens/s
345
ms
pp2048+tg768
6084
tokens/s
120
tokens/s
345
ms
pp1024+tg1024
6367
tokens/s
123
tokens/s
169
ms
pp1280+tg3072
6157
tokens/s
118
tokens/s
216
ms
pp384+tg1152
6223
tokens/s
126
tokens/s
70
ms
pp64+tg1024
3070
tokens/s
128
tokens/s
29
ms
pp16+tg1536
1121
tokens/s
126
tokens/s
22
ms