TEST #864 RESULTS
05/14/2025 - 8:22 AM
ACCELERATOR
MODEL
21.5
tokens/s
generation
5.65
sec
time to first token
223
tokens/s
prompt
95
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 Pro 5P+6E
RAM
36GB
OS
Darwin
Kernel Release
24.5.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.5.0: Tue Apr 22 19:54:29 PDT 2025; root:xnu-11417.121.6~2/RELEASE_ARM64_T6030
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
238
tokens/s
17.0
tokens/s
4.37
sec
pp4096+tg256
222
tokens/s
18.2
tokens/s
18.53
sec
pp2048+tg256
247
tokens/s
21.4
tokens/s
8.34
sec
pp2048+tg768
247
tokens/s
20.9
tokens/s
8.35
sec
pp1024+tg1024
255
tokens/s
22.6
tokens/s
4.06
sec
pp1280+tg3072
253
tokens/s
20.3
tokens/s
5.10
sec
pp384+tg1152
255
tokens/s
23.8
tokens/s
1.54
sec
pp64+tg1024
209
tokens/s
24.7
tokens/s
345
ms
pp16+tg1536
85
tokens/s
24.1
tokens/s
226
ms