The paper introduces PhysicsAssistant, an interactive learning robot designed to assist students in physics lab investigations. PhysicsAssistant integrates YOLOv8 for object detection, GPT-3.5-turbo for natural language processing, and speech recognition to provide real-time, multimodal support. A user study involving ten 8th-grade students evaluated the system's performance using Bloom's taxonomy, comparing it with GPT-4. The results show that PhysicsAssistant, while slightly less effective in conceptual and procedural knowledge, is significantly faster in response time compared to GPT-4. Despite these limitations, PhysicsAssistant demonstrates potential as a real-time lab assistant, offering timely responses and offloading teacher labor. The system's ability to handle multimodal input and provide contextually relevant responses makes it a promising tool for enhancing K-12 physics education.The paper introduces PhysicsAssistant, an interactive learning robot designed to assist students in physics lab investigations. PhysicsAssistant integrates YOLOv8 for object detection, GPT-3.5-turbo for natural language processing, and speech recognition to provide real-time, multimodal support. A user study involving ten 8th-grade students evaluated the system's performance using Bloom's taxonomy, comparing it with GPT-4. The results show that PhysicsAssistant, while slightly less effective in conceptual and procedural knowledge, is significantly faster in response time compared to GPT-4. Despite these limitations, PhysicsAssistant demonstrates potential as a real-time lab assistant, offering timely responses and offloading teacher labor. The system's ability to handle multimodal input and provide contextually relevant responses makes it a promising tool for enhancing K-12 physics education.