TEST #1752 RESULTS
11/08/2025 - 12:54 AM
ACCELERATOR
MODEL
16.9
tokens/s
generation
7.44
sec
time to first token
173
tokens/s
prompt
73
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M4 4P+6E
RAM
16GB
OS
Darwin
Kernel Release
25.1.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.1.0: Mon Oct 20 19:32:56 PDT 2025; root:xnu-12377.41.6~2/RELEASE_ARM64_T8132
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
203
tokens/s
17.9
tokens/s
5.10
sec
pp4096+tg256
173
tokens/s
14.6
tokens/s
23.80
sec
pp2048+tg256
184
tokens/s
16.3
tokens/s
11.20
sec
pp2048+tg768
179
tokens/s
15.9
tokens/s
11.48
sec
pp1024+tg1024
182
tokens/s
17.3
tokens/s
5.69
sec
pp1280+tg3072
184
tokens/s
15.6
tokens/s
7.03
sec
pp384+tg1152
195
tokens/s
18.0
tokens/s
2.03
sec
pp64+tg1024
183
tokens/s
18.4
tokens/s
400
ms
pp16+tg1536
79
tokens/s
18.1
tokens/s
255
ms