TEST #2556 RESULTS
02/07/2026 - 5:06 AM
ACCELERATOR
MODEL
31.2
tokens/s
generation
3.77
sec
time to first token
325
tokens/s
prompt
139
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M2 Max 8P+4E
RAM
32GB
OS
Darwin
Kernel Release
24.6.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.6.0: Wed Nov 5 21:32:34 PST 2025; root:xnu-11417.140.69.705.2~1/RELEASE_ARM64_T6020
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
460
tokens/s
43.3
tokens/s
2.25
sec
pp4096+tg256
348
tokens/s
26.8
tokens/s
11.80
sec
pp2048+tg256
358
tokens/s
30.8
tokens/s
5.75
sec
pp2048+tg768
357
tokens/s
29.7
tokens/s
5.76
sec
pp1024+tg1024
363
tokens/s
29.4
tokens/s
2.85
sec
pp1280+tg3072
334
tokens/s
25.8
tokens/s
3.87
sec
pp384+tg1152
333
tokens/s
30.8
tokens/s
1.18
sec
pp64+tg1024
268
tokens/s
31.5
tokens/s
263
ms
pp16+tg1536
103
tokens/s
32.5
tokens/s
178
ms