Text this: Benchmarking State-of-the-Art Large Language Models for Migraine Patient Education: Performance Comparison of Responses to Common Queries