Add option for Gemini 0605 to enable reasoning but not set manual max-output-tokens and max-thinking-tokens

[ParthK117]

I notice that the new Gemini model is given the same options as other thinking models in setting enabling/disabling its reasoning ability and setting max tokens.

However, the new gemini model is unique in that thinking cannot be turned off, in fact the option is grayed out in th AI Studio too.

What can be done however is either set the reasoning budget yourself (the option available in roo) or like 0506, let the model determine the optimal thinking budget for each request.

It would be a simple fix to rename the 'Enable reasoning' checkbox to a better option and function of 'Set Reasoning Budget' which when checked show the same reasoning budget sliders as they do now, (allowing someone to set reasoning to the minimum of 128 if they so choose) and unchecking it allowing the model to determine its own budget.