TEST #1361 RESULTS
08/12/2025 - 12:18 AM
ACCELERATOR
MODEL
63.3
tokens/s
generation
1.10
sec
time to first token
1062
tokens/s
prompt
394
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 Ultra 24P+8E
RAM
256GB
OS
Darwin
Kernel Release
24.6.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.6.0: Thu Jul 31 21:25:54 PDT 2025; root:xnu-11417.140.69.701.6~1/RELEASE_ARM64_T6031
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1288
tokens/s
68.3
tokens/s
809
ms
pp4096+tg256
1196
tokens/s
48.6
tokens/s
3.45
sec
pp2048+tg256
1271
tokens/s
59.5
tokens/s
1.63
sec
pp2048+tg768
1270
tokens/s
58.0
tokens/s
1.63
sec
pp1024+tg1024
1278
tokens/s
63.8
tokens/s
816
ms
pp1280+tg3072
1252
tokens/s
56.3
tokens/s
1.04
sec
pp384+tg1152
1209
tokens/s
69.6
tokens/s
331
ms
pp64+tg1024
618
tokens/s
73.5
tokens/s
117
ms
pp16+tg1536
178
tokens/s
72.1
tokens/s
103
ms