TEST #2981 RESULTS
03/05/2026 - 8:55 PM
ACCELERATOR
MODEL
15.6
tokens/s
generation
8.17
sec
time to first token
157
tokens/s
prompt
67
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 4P+4E
RAM
16GB
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 20:53:31 PST 2026; root:xnu-12377.81.4~5/RELEASE_ARM64_T8122
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
172
tokens/s
16.4
tokens/s
6.00
sec
pp4096+tg256
153
tokens/s
13.3
tokens/s
26.76
sec
pp2048+tg256
168
tokens/s
14.6
tokens/s
12.25
sec
pp2048+tg768
167
tokens/s
15.0
tokens/s
12.29
sec
pp1024+tg1024
176
tokens/s
15.8
tokens/s
5.88
sec
pp1280+tg3072
174
tokens/s
14.7
tokens/s
7.41
sec
pp384+tg1152
178
tokens/s
16.8
tokens/s
2.22
sec
pp64+tg1024
161
tokens/s
16.7
tokens/s
453
ms
pp16+tg1536
66
tokens/s
17.0
tokens/s
298
ms