TEST #1190 RESULTS
07/07/2025 - 11:36 AM
ACCELERATOR
MODEL
54.2
tokens/s
generation
1.80
sec
time to first token
663
tokens/s
prompt
271
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M4 Max 12P+4E
RAM
48GB
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 19:53:27 PDT 2025; root:xnu-11417.121.6~2/RELEASE_ARM64_T6041
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
815
tokens/s
60.8
tokens/s
1.27
sec
pp4096+tg256
751
tokens/s
43.2
tokens/s
5.48
sec
pp2048+tg256
782
tokens/s
53.0
tokens/s
2.64
sec
pp2048+tg768
782
tokens/s
51.5
tokens/s
2.64
sec
pp1024+tg1024
731
tokens/s
56.6
tokens/s
1.42
sec
pp1280+tg3072
661
tokens/s
49.4
tokens/s
1.95
sec
pp384+tg1152
735
tokens/s
58.9
tokens/s
538
ms
pp64+tg1024
529
tokens/s
59.5
tokens/s
136
ms
pp16+tg1536
177
tokens/s
54.8
tokens/s
105
ms