TEST #2666 RESULTS
02/15/2026 - 5:39 PM
ACCELERATOR
MODEL
38.4
tokens/s
generation
3.08
sec
time to first token
392
tokens/s
prompt
170
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
24.6.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.6.0: Mon Jul 14 11:30:29 PDT 2025; root:xnu-11417.140.69~1/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
41.5
tokens/s
2.29
sec
pp4096+tg256
425
tokens/s
30.2
tokens/s
9.66
sec
pp2048+tg256
443
tokens/s
36.4
tokens/s
4.65
sec
pp2048+tg768
445
tokens/s
35.8
tokens/s
4.63
sec
pp1024+tg1024
451
tokens/s
39.2
tokens/s
2.29
sec
pp1280+tg3072
448
tokens/s
34.5
tokens/s
2.88
sec
pp384+tg1152
442
tokens/s
41.8
tokens/s
891
ms
pp64+tg1024
317
tokens/s
43.6
tokens/s
224
ms
pp16+tg1536
105
tokens/s
42.6
tokens/s
173
ms