the fact that openai hides the reasoning tokens from us to begin with shows that what they are doing behind the scenes isnt all that impressive, and likely easily cloned (r1)

would be nice if they made them visible now