TEST #579 RESULTS
04/08/2025 - 10:12 AM
ACCELERATOR
MODEL
22.1
tokens/s
generation
4.46
sec
time to first token
284
tokens/s
prompt
112
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Apple M3 Pro 6P+6E
RAM
36GB
OS
Darwin
Kernel Release
24.3.0
Architecture
arm64
Version
Cosmopolitan 3.9.7 MODE=aarch64; Darwin Kernel Version 24.3.0: Thu Jan 2 20:24:24 PST 2025; root:xnu-11215.81.4~3/RELEASE_ARM64_T6030
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
324
tokens/s
23.6
tokens/s
3.20
sec
pp4096+tg256
281
tokens/s
18.3
tokens/s
14.62
sec
pp2048+tg256
310
tokens/s
21.4
tokens/s
6.66
sec
pp2048+tg768
310
tokens/s
20.9
tokens/s
6.66
sec
pp1024+tg1024
321
tokens/s
22.5
tokens/s
3.23
sec
pp1280+tg3072
319
tokens/s
20.2
tokens/s
4.05
sec
pp384+tg1152
318
tokens/s
23.1
tokens/s
1.25
sec
pp64+tg1024
261
tokens/s
24.4
tokens/s
285
ms
pp16+tg1536
110
tokens/s
24.0
tokens/s
184
ms