potentially, I can achieve it by setting maximum tokens ai can use for response. For example, on code in this [pr](https://github.com/CodeYourFuture/Module-Tools-Mirror-For-AI-Code-Reviewer-Testing/pull/2) Ai spend: - around 2,500 tokens to give ≈19 feedback points without reasoning, - 4,000 tokens to give 6 comments with medium reasoning
potentially, I can achieve it by setting maximum tokens ai can use for response.
For example, on code in this pr Ai spend: