TEST #3293 RESULTS
04/12/2026 - 12:40 AM
ACCELERATOR
MODEL
17.4
tokens/s
generation
6.99
sec
time to first token
180
tokens/s
prompt
77
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
209
tokens/s
18.5
tokens/s
4.95
sec
pp4096+tg256
184
tokens/s
15.0
tokens/s
22.27
sec
pp2048+tg256
200
tokens/s
16.9
tokens/s
10.28
sec
pp2048+tg768
192
tokens/s
16.6
tokens/s
10.74
sec
pp1024+tg1024
195
tokens/s
17.7
tokens/s
5.29
sec
pp1280+tg3072
193
tokens/s
16.1
tokens/s
6.67
sec
pp384+tg1152
198
tokens/s
18.3
tokens/s
1.99
sec
pp64+tg1024
175
tokens/s
18.9
tokens/s
417
ms
pp16+tg1536
76
tokens/s
18.6
tokens/s
262
ms