as an avid prompter for over 2 years now. I can say it feels alot worse to have a model think for you when you have prompts that can out think it. It cause alot of uncessesarry interference.
If this is just for testing a Reasoning model or makes sense though, which is why i assume you removed 2.0 pro so you can test the Thinking then leaving it makes sense.
But i will say having it be Optional for all future models that people used would be a huge step forward considering alot of the times the prompt is direct correlation with the models output, which hogh level prompters can leverage to get the most out of these models and with Reasoning forces youre nerfing your models potential intelligence ironically.