I said it a thousand times, general AI cannot work without logic. Maybe those fiat assholes finally got the memo.

It's on like o3 right now, probably because the logic, and reasoning aspects are still early, and need to be fleshed out.

I'm sure the primary concern is model run time, usually they limit most models to like a minute at most to prevent rogue thinking.

I'm not sure but the more time an AI is given to think the better its answers tend to be.