TEST #790 RESULTS
05/01/2025 - 6:36 AM
ACCELERATOR
MODEL
37.7
tokens/s
generation
2.69
sec
time to first token
457
tokens/s
prompt
186
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 Max 10P+4E
RAM
96GB
OS
Darwin
Kernel Release
24.4.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.4.0: Fri Apr 11 18:33:40 PDT 2025; root:xnu-11417.101.15~117/RELEASE_ARM64_T6031
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
524
tokens/s
40.7
tokens/s
1.98
sec
pp4096+tg256
481
tokens/s
30.0
tokens/s
8.55
sec
pp2048+tg256
511
tokens/s
36.1
tokens/s
4.03
sec
pp2048+tg768
511
tokens/s
35.2
tokens/s
4.03
sec
pp1024+tg1024
524
tokens/s
38.5
tokens/s
1.98
sec
pp1280+tg3072
512
tokens/s
33.8
tokens/s
2.52
sec
pp384+tg1152
513
tokens/s
41.0
tokens/s
772
ms
pp64+tg1024
390
tokens/s
42.7
tokens/s
186
ms
pp16+tg1536
145
tokens/s
41.7
tokens/s
132
ms