TEST #3133 RESULTS
03/20/2026 - 2:00 PM
ACCELERATOR
MODEL
18.1
tokens/s
generation
6.90
sec
time to first token
185
tokens/s
prompt
79
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
25.3.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.3.0: Wed Jan 28 20:49:24 PST 2026; root:xnu-12377.81.4~5/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
18.9
tokens/s
4.94
sec
pp4096+tg256
185
tokens/s
15.4
tokens/s
22.17
sec
pp2048+tg256
202
tokens/s
17.6
tokens/s
10.20
sec
pp2048+tg768
197
tokens/s
17.3
tokens/s
10.43
sec
pp1024+tg1024
199
tokens/s
18.4
tokens/s
5.19
sec
pp1280+tg3072
195
tokens/s
16.8
tokens/s
6.63
sec
pp384+tg1152
202
tokens/s
19.2
tokens/s
1.95
sec
pp64+tg1024
193
tokens/s
19.7
tokens/s
382
ms
pp16+tg1536
85
tokens/s
19.4
tokens/s
238
ms