TEST #1954 RESULTS
12/16/2025 - 9:02 PM
ACCELERATOR
MODEL
66.2
tokens/s
generation
1.44
sec
time to first token
895
tokens/s
prompt
345
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M4 4P+6E
RAM
16GB
OS
Darwin
Kernel Release
25.2.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.2.0: Tue Nov 18 21:08:48 PST 2025; root:xnu-12377.61.12~1/RELEASE_ARM64_T8132
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1046
tokens/s
74.0
tokens/s
993
ms
pp4096+tg256
827
tokens/s
46.5
tokens/s
4.97
sec
pp2048+tg256
979
tokens/s
61.0
tokens/s
2.11
sec
pp2048+tg768
978
tokens/s
58.8
tokens/s
2.11
sec
pp1024+tg1024
1043
tokens/s
67.6
tokens/s
995
ms
pp1280+tg3072
1033
tokens/s
55.4
tokens/s
1.25
sec
pp384+tg1152
1049
tokens/s
74.5
tokens/s
378
ms
pp64+tg1024
865
tokens/s
80.2
tokens/s
85
ms
pp16+tg1536
235
tokens/s
77.3
tokens/s
79
ms