Because the environment did not reward being as efficient as other models?
Oh so it's using more tokens to do the same work. Make sense, thanks!
Oh so it's using more tokens to do the same work. Make sense, thanks!