TEST #3537 RESULTS
05/15/2026 - 4:46 AM
ACCELERATOR
MODEL
17.1
tokens/s
generation
7.15
sec
time to first token
178
tokens/s
prompt
75
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
205
tokens/s
18.3
tokens/s
5.04
sec
pp4096+tg256
182
tokens/s
14.9
tokens/s
22.62
sec
pp2048+tg256
193
tokens/s
17.0
tokens/s
10.68
sec
pp2048+tg768
186
tokens/s
16.3
tokens/s
11.05
sec
pp1024+tg1024
192
tokens/s
17.4
tokens/s
5.40
sec
pp1280+tg3072
189
tokens/s
15.8
tokens/s
6.83
sec
pp384+tg1152
193
tokens/s
17.1
tokens/s
2.05
sec
pp64+tg1024
183
tokens/s
18.8
tokens/s
400
ms
pp16+tg1536
79
tokens/s
18.4
tokens/s
252
ms