4 ms·still quite verbose at 140m output tokens, but this is on max thinking. high should do better.by sourcecodeplz 3mo agostill quite verbose at 140m output tokens, but this is on max thinking. high should do better.