TEST #804 RESULTS
05/03/2025 - 9:43 PM
ACCELERATOR
MODEL
25.4
tokens/s
generation
5.68
sec
time to first token
256
tokens/s
prompt
105
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 3700X 8-Core Processor (znver2)
RAM
32GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
284
tokens/s
31.4
tokens/s
3.63
sec
pp4096+tg256
191
tokens/s
9.8
tokens/s
21.51
sec
pp2048+tg256
259
tokens/s
22.6
tokens/s
7.95
sec
pp2048+tg768
259
tokens/s
20.6
tokens/s
7.97
sec
pp1024+tg1024
280
tokens/s
26.2
tokens/s
3.69
sec
pp1280+tg3072
275
tokens/s
16.7
tokens/s
4.69
sec
pp384+tg1152
295
tokens/s
31.3
tokens/s
1.33
sec
pp64+tg1024
249
tokens/s
35.7
tokens/s
281
ms
pp16+tg1536
215
tokens/s
33.9
tokens/s
98
ms