TEST #11 RESULTS
03/19/2025 - 10:30 PM
ACCELERATOR
MODEL
51.6
tokens/s
generation
1.99
sec
time to first token
607
tokens/s
prompt
250
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M4 Max 12P+4E
RAM
128GB
OS
Darwin
Kernel Release
24.3.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.3.0: Thu Jan 2 20:24:22 PST 2025; root:xnu-11215.81.4~3/RELEASE_ARM64_T6041
RUNTIME
Name
llamafile
Version
0.9.1
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
817
tokens/s
61.4
tokens/s
1.27
sec
pp4096+tg256
689
tokens/s
41.7
tokens/s
5.97
sec
pp2048+tg256
661
tokens/s
50.7
tokens/s
3.12
sec
pp2048+tg768
656
tokens/s
49.2
tokens/s
3.14
sec
pp1024+tg1024
682
tokens/s
53.6
tokens/s
1.52
sec
pp1280+tg3072
621
tokens/s
46.8
tokens/s
2.08
sec
pp384+tg1152
662
tokens/s
55.6
tokens/s
597
ms
pp64+tg1024
510
tokens/s
55.5
tokens/s
141
ms
pp16+tg1536
169
tokens/s
49.6
tokens/s
110
ms