O1 and O1-mini models

Hi, why did you shorten the context window for the o1 and o1-mini models? It’s giving really short answers, like just a few lines, and then it just stops mid-sentence. This model is usually good at giving more text, but you’ve really cut it down a lot.