TEST #4008 RESULTS
08/09/2026 - 7:36 PM
ACCELERATOR
MODEL
173
tokens/s
generation
261
ms
time to first token
4394
tokens/s
prompt
1428
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 Ultra 20P+8E
RAM
96GB
OS
Darwin
Kernel Release
25.6.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.6.0: Sat Jul 11 15:27:26 PDT 2026; root:xnu-12377.161.13~4/RELEASE_ARM64_T6031
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
5775
tokens/s
191
tokens/s
183
ms
pp4096+tg256
5093
tokens/s
126
tokens/s
812
ms
pp2048+tg256
5607
tokens/s
161
tokens/s
372
ms
pp2048+tg768
5502
tokens/s
156
tokens/s
379
ms
pp1024+tg1024
5467
tokens/s
176
tokens/s
193
ms
pp1280+tg3072
5433
tokens/s
147
tokens/s
241
ms
pp384+tg1152
4546
tokens/s
192
tokens/s
89
ms
pp64+tg1024
1617
tokens/s
208
tokens/s
44
ms
pp16+tg1536
505
tokens/s
200
tokens/s
36
ms