TEST #1628 RESULTS
10/09/2025 - 3:19 PM
ACCELERATOR
MODEL
54.4
tokens/s
generation
1.67
sec
time to first token
675
tokens/s
prompt
280
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M1 Ultra 16P+4E
RAM
128GB
OS
Darwin
Kernel Release
25.0.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.0.0: Wed Sep 17 21:41:45 PDT 2025; root:xnu-12377.1.9~141/RELEASE_ARM64_T6000
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
831
tokens/s
58.0
tokens/s
1.25
sec
pp4096+tg256
811
tokens/s
45.0
tokens/s
5.07
sec
pp2048+tg256
838
tokens/s
52.2
tokens/s
2.46
sec
pp2048+tg768
841
tokens/s
51.6
tokens/s
2.45
sec
pp1024+tg1024
825
tokens/s
55.1
tokens/s
1.26
sec
pp1280+tg3072
811
tokens/s
50.0
tokens/s
1.59
sec
pp384+tg1152
704
tokens/s
58.2
tokens/s
562
ms
pp64+tg1024
318
tokens/s
60.9
tokens/s
217
ms
pp16+tg1536
97
tokens/s
59.0
tokens/s
180
ms