TEST #3390 RESULTS
04/26/2026 - 4:52 PM
ACCELERATOR
MODEL
16.4
tokens/s
generation
7.42
sec
time to first token
172
tokens/s
prompt
72
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
25.2.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.2.0: Tue Nov 18 21:08:48 PST 2025; root:xnu-12377.61.12~1/RELEASE_ARM64_T8132
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
196
tokens/s
17.7
tokens/s
5.28
sec
pp4096+tg256
177
tokens/s
13.2
tokens/s
23.16
sec
pp2048+tg256
175
tokens/s
14.3
tokens/s
11.76
sec
pp2048+tg768
180
tokens/s
14.3
tokens/s
11.44
sec
pp1024+tg1024
183
tokens/s
17.1
tokens/s
5.67
sec
pp1280+tg3072
189
tokens/s
15.4
tokens/s
6.82
sec
pp384+tg1152
197
tokens/s
18.2
tokens/s
2.01
sec
pp64+tg1024
176
tokens/s
18.8
tokens/s
414
ms
pp16+tg1536
78
tokens/s
18.6
tokens/s
255
ms