TEST #1552 RESULTS
09/24/2025 - 9:10 AM
ACCELERATOR
MODEL
133
tokens/s
generation
467
ms
time to first token
2553
tokens/s
prompt
899
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 Max 10P+4E
RAM
36GB
OS
Darwin
Kernel Release
25.0.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.0.0: Mon Aug 25 21:16:39 PDT 2025; root:xnu-12377.1.9~3/RELEASE_ARM64_T6031
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
3231
tokens/s
148
tokens/s
324
ms
pp4096+tg256
2692
tokens/s
93.1
tokens/s
1.53
sec
pp2048+tg256
3021
tokens/s
122
tokens/s
686
ms
pp2048+tg768
3092
tokens/s
118
tokens/s
670
ms
pp1024+tg1024
3164
tokens/s
136
tokens/s
330
ms
pp1280+tg3072
3103
tokens/s
111
tokens/s
419
ms
pp384+tg1152
2865
tokens/s
150
tokens/s
140
ms
pp64+tg1024
1462
tokens/s
162
tokens/s
49
ms
pp16+tg1536
350
tokens/s
155
tokens/s
51
ms