Text this: Comparative Evaluation of Multimodal Large Language Models for No-Reference Image Quality Assessment with Authentic Distortions: A Study of OpenAI and Claude.AI Models