Implement Query Response Rating System

Currently there is no structured way for users to rate the quality of AI responses. This proposal introduces a simple thumbs-up/thumbs-down rating system on each model's response, with optional written feedback. Aggregated ratings would help identify which models perform best for different query types, inform the community's model preferences, and provide valuable data for improving the platform's routing logic.