Diagnostic Performance of ChatGPT-o1 and DeepSeek-V3 in Expert-Validated Simulated Ear Nose and Throat Scenarios: A Comparative Accuracy Study | Synapse