The exercise involved five AI models responding to a prompt to create a Rust function for determining leap years. Performance varied, with Kilo Code models demonstrating detailed explanations and accuracy. ChatGPT and Claude excelled in using the Rust test harness, while Qwen asked clarifying questions. Overall, the models were effective but had differing strengths.