A comparison of the diagnostic ability of large language models in challenging clinical cases | Synapse