alt
Hacker News
ig0r0
•
yesterday at 8:48 PM
•
1 reply
•
view on HN
yes, with qwen3.8-27b-4bit run via rapid-mlx i can get to about 200k
Replies
metadat
•
yesterday at 9:03 PM
Nice.. and how many toks/sec?
➕ show 1 reply
Nice.. and how many toks/sec?