OneTutor is designed as a course-specific learning companion, not as a general-purpose chatbot. The quality of its answers therefore depends not only on the underlying AI model, but also on the course context provided.
Instructors can upload their own materials and define the subject-specific basis on which the tutor responds. They can also control how strictly the tutor should follow the provided materials. This makes it possible to align answers more closely with the lecture, its terminology, and the intended definitions.
Still, as with all AI systems, mistakes can occur. This is especially relevant for complex proofs, highly specific definitions, or multi-step calculations. OneTutor should therefore be used as a tool for understanding, review, and practice, not as a replacement for expert review or official model solutions.