here is o1, a series of our most capable and aligned models yet: https://
openai.com/index/learning
-to-reason-with-llms/
… o1 is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it.

→ View original post on X — @sama,