TEST #1749 RESULTS
11/07/2025 - 10:53 PM
ACCELERATOR
MODEL
20.8
tokens/s
generation
4.53
sec
time to first token
279
tokens/s
prompt
109
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 Pro 6P+6E
RAM
18GB
OS
Darwin
Kernel Release
24.6.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.6.0: Mon Aug 11 21:16:31 PDT 2025; root:xnu-11417.140.69.701.11~1/RELEASE_ARM64_T6030
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
322
tokens/s
21.3
tokens/s
3.23
sec
pp4096+tg256
275
tokens/s
17.1
tokens/s
14.93
sec
pp2048+tg256
307
tokens/s
19.5
tokens/s
6.73
sec
pp2048+tg768
305
tokens/s
19.9
tokens/s
6.77
sec
pp1024+tg1024
316
tokens/s
21.2
tokens/s
3.28
sec
pp1280+tg3072
317
tokens/s
19.3
tokens/s
4.08
sec
pp384+tg1152
323
tokens/s
22.6
tokens/s
1.23
sec
pp64+tg1024
265
tokens/s
23.6
tokens/s
281
ms
pp16+tg1536
83
tokens/s
22.7
tokens/s
237
ms