Upgraded the AI infrastructure across the application by transitioning from gemini-2.5-flash to the newer gemini-3.5-flash model. This update applies to all content generation, automated grading, and vision tasks, ensuring the service benefits from the latest model capabilities.