Effective strategies to control model cost

You are developing a prototype of an AI-powered code generation tool for your company, using a commercial LLM. During a review meeting, your manager expresses concern about the long-term costs of running the tool. They are particularly worried about how quickly token usage could grow as more developers begin using it, and how unpredictable the expenses might become.

To move forward responsibly, identify which of the following options is the most effective strategy to control token cost while maintaining acceptable performance.