TEST #3131 RESULTS
03/20/2026 - 1:38 PM
ACCELERATOR
MODEL
21.6
tokens/s
generation
6.86
sec
time to first token
187
tokens/s
prompt
84
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M2 4P+4E
RAM
8GB
OS
Darwin
Kernel Release
25.3.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.3.0: Wed Jan 28 20:56:34 PST 2026; root:xnu-12377.91.3~2/RELEASE_ARM64_T8112
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
207
tokens/s
23.2
tokens/s
4.99
sec
pp4096+tg256
184
tokens/s
16.8
tokens/s
22.31
sec
pp2048+tg256
200
tokens/s
20.6
tokens/s
10.29
sec
pp2048+tg768
199
tokens/s
20.1
tokens/s
10.33
sec
pp1024+tg1024
207
tokens/s
22.2
tokens/s
4.98
sec
pp1280+tg3072
204
tokens/s
19.1
tokens/s
6.33
sec
pp384+tg1152
210
tokens/s
23.5
tokens/s
1.87
sec
pp64+tg1024
192
tokens/s
24.9
tokens/s
371
ms
pp16+tg1536
80
tokens/s
24.2
tokens/s
240
ms