TEST #2915 RESULTS
02/27/2026 - 4:00 AM
ACCELERATOR
MODEL
37.5
tokens/s
generation
3.17
sec
time to first token
369
tokens/s
prompt
164
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M1 Max 8P+2E
RAM
64GB
OS
Darwin
Kernel Release
25.2.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 25.2.0: Tue Nov 18 21:09:40 PST 2025; root:xnu-12377.61.12~1/RELEASE_ARM64_T6000
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
434
tokens/s
39.8
tokens/s
2.38
sec
pp4096+tg256
418
tokens/s
29.8
tokens/s
9.83
sec
pp2048+tg256
439
tokens/s
36.1
tokens/s
4.70
sec
pp2048+tg768
438
tokens/s
35.1
tokens/s
4.70
sec
pp1024+tg1024
443
tokens/s
38.3
tokens/s
2.34
sec
pp1280+tg3072
422
tokens/s
33.8
tokens/s
3.06
sec
pp384+tg1152
395
tokens/s
40.5
tokens/s
997
ms
pp64+tg1024
243
tokens/s
42.6
tokens/s
285
ms
pp16+tg1536
89
tokens/s
41.7
tokens/s
202
ms