TEST #899 RESULTS
05/18/2025 - 7:12 PM
ACCELERATOR
MODEL
33.7
tokens/s
generation
3.54
sec
time to first token
350
tokens/s
prompt
149
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M1 Max 8P+2E
RAM
32GB
OS
Darwin
Kernel Release
24.5.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.5.0: Tue Apr 22 19:54:49 PDT 2025; root:xnu-11417.121.6~2/RELEASE_ARM64_T6000
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
454
tokens/s
40.7
tokens/s
2.28
sec
pp4096+tg256
371
tokens/s
28.4
tokens/s
11.08
sec
pp2048+tg256
381
tokens/s
32.7
tokens/s
5.40
sec
pp2048+tg768
375
tokens/s
31.4
tokens/s
5.50
sec
pp1024+tg1024
388
tokens/s
33.2
tokens/s
2.66
sec
pp1280+tg3072
364
tokens/s
29.9
tokens/s
3.55
sec
pp384+tg1152
395
tokens/s
34.9
tokens/s
998
ms
pp64+tg1024
313
tokens/s
36.2
tokens/s
227
ms
pp16+tg1536
111
tokens/s
35.4
tokens/s
166
ms