TEST #3374 RESULTS
04/23/2026 - 7:53 PM
ACCELERATOR
MODEL
32.3
tokens/s
generation
3.70
sec
time to first token
338
tokens/s
prompt
143
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M1 Max 8P+2E
RAM
64GB
OS
Darwin
Kernel Release
25.4.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.4.0: Thu Mar 19 19:30:44 PDT 2026; root:xnu-12377.101.15~1/RELEASE_ARM64_T6000
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
451
tokens/s
40.7
tokens/s
2.29
sec
pp4096+tg256
359
tokens/s
27.2
tokens/s
11.45
sec
pp2048+tg256
363
tokens/s
31.2
tokens/s
5.67
sec
pp2048+tg768
354
tokens/s
29.9
tokens/s
5.82
sec
pp1024+tg1024
346
tokens/s
31.4
tokens/s
3.00
sec
pp1280+tg3072
344
tokens/s
27.3
tokens/s
3.75
sec
pp384+tg1152
408
tokens/s
33.4
tokens/s
966
ms
pp64+tg1024
309
tokens/s
34.8
tokens/s
229
ms
pp16+tg1536
107
tokens/s
34.3
tokens/s
171
ms