Y
Hacker News
new
|
ask
|
show
|
jobs
by
sourcecodeplz
6 hours ago
still quite verbose at 140m output tokens, but this is on max thinking. high should do better.