Streaming Reasoning Output

It seems like reasoning is not viewable until after the response is complete. I would like the ability to view the reasoning as it streams, especially for slower local models.