TEST #1104 RESULTS
06/24/2025 - 9:40 PM
ACCELERATOR
MODEL
9.7
tokens/s
generation
13.99
sec
time to first token
99
tokens/s
prompt
41
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 4P+4E
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 19:54:33 PDT 2025; root:xnu-11417.121.6~2/RELEASE_ARM64_T8122
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
166
tokens/s
15.1
tokens/s
6.25
sec
pp4096+tg256
133
tokens/s
9.5
tokens/s
30.80
sec
pp2048+tg256
62
tokens/s
4.5
tokens/s
33.03
sec
pp2048+tg768
82
tokens/s
5.4
tokens/s
25.15
sec
pp1024+tg1024
90
tokens/s
5.5
tokens/s
11.41
sec
pp1280+tg3072
89
tokens/s
6.6
tokens/s
14.53
sec
pp384+tg1152
109
tokens/s
9.4
tokens/s
3.61
sec
pp64+tg1024
87
tokens/s
15.9
tokens/s
809
ms
pp16+tg1536
69
tokens/s
15.3
tokens/s
290
ms