Visual Studio 18.9 Insiders 2 introduces a thinking effort control for AI models in the IDE. Users can now set reasoning levels — Low, Medium, High, Extra High, and Max — per model, matching the depth of AI reasoning to the complexity of the task. Simple tasks like renaming variables can use Low effort (fewer credits consumed), while complex debugging or architecture decisions can use High or Max. The setting lives in the model management view alongside model capabilities, context size, and cost, making it easy to tune in one place.
Table of contents
A dial, not a switch Copy linkSet it where you already work Copy linkConsume your usage credits where they count Copy linkGive it a try Copy link436 Impressions