TEST #1085 RESULTS
06/20/2025 - 9:43 PM
ACCELERATOR
MODEL
16.6
tokens/s
generation
7.93
sec
time to first token
159
tokens/s
prompt
69
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M4 4P+6E
RAM
24GB
OS
Darwin
Kernel Release
24.5.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.5.0: Tue Apr 22 19:54:43 PDT 2025; root:xnu-11417.121.6~2/RELEASE_ARM64_T8132
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
183
tokens/s
18.9
tokens/s
5.66
sec
pp4096+tg256
160
tokens/s
14.5
tokens/s
25.63
sec
pp2048+tg256
176
tokens/s
17.3
tokens/s
11.71
sec
pp2048+tg768
171
tokens/s
16.8
tokens/s
12.02
sec
pp1024+tg1024
169
tokens/s
17.2
tokens/s
6.12
sec
pp1280+tg3072
179
tokens/s
16.0
tokens/s
7.19
sec
pp384+tg1152
174
tokens/s
18.1
tokens/s
2.26
sec
pp64+tg1024
159
tokens/s
16.9
tokens/s
455
ms
pp16+tg1536
58
tokens/s
13.4
tokens/s
331
ms