Example of how to test and measure prompt performance across different models - View it on GitHub