TEST #3303 RESULTS
04/14/2026 - 3:11 AM
ACCELERATOR
MODEL
37.7
tokens/s
generation
3.09
sec
time to first token
389
tokens/s
prompt
168
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
25.3.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.3.0: Wed Jan 28 20:54:46 PST 2026; root:xnu-12377.91.3~2/RELEASE_ARM64_T6000
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
451
tokens/s
40.7
tokens/s
2.29
sec
pp4096+tg256
424
tokens/s
30.1
tokens/s
9.69
sec
pp2048+tg256
445
tokens/s
36.1
tokens/s
4.63
sec
pp2048+tg768
444
tokens/s
35.3
tokens/s
4.64
sec
pp1024+tg1024
450
tokens/s
38.5
tokens/s
2.30
sec
pp1280+tg3072
445
tokens/s
34.0
tokens/s
2.90
sec
pp384+tg1152
435
tokens/s
40.9
tokens/s
905
ms
pp64+tg1024
302
tokens/s
42.7
tokens/s
234
ms
pp16+tg1536
101
tokens/s
40.5
tokens/s
181
ms