ChatGPT is today a better doctor than your doctor, with almost a hundred percent certainty.

The test is whether, as of October 2025, the evidence shows ChatGPT performing better than typical practising physicians at real clinical care in general (diagnosis, treatment decisions, patient management), not only on narrow tasks. True: peer-reviewed studies or systematic reviews show ChatGPT broadly outperforming typical physicians across clinical practice, including in real-world or prospective settings. Mostly true: strong, consistent evidence that ChatGPT beats typical physicians across most major clinical tasks, with minor exceptions. Mixed: the evidence is roughly split, with ChatGPT outperforming on a substantial share of clinically meaningful tasks and underperforming on a comparable share. Mostly false: ChatGPT outperforms physicians only on narrow or artificial measures (exam questions, curated case vignettes, written empathy ratings), or real-world evidence is lacking, so the general claim is unproven or overstated. False: the evidence shows physicians generally match or outperform ChatGPT on clinical care.

Set before the research began.

1. E1 A systematic review and meta-analysis of diagnostic performance comparison between generative AI and physicians - npj Digital Medicine npj Digital Medicine (Springer Nature) · 22 March 2025

   “No significant performance difference was found between AI models and physicians overall (p = 0.10) or non-expert physicians (p = 0.93). However, AI models performed significantly worse than expert physicians (p = 0.007).”

2. E2 ChatGPT (GPT-4) versus doctors on complex cases of the Swedish family medicine specialist examination: an observational comparative study PubMed Central (NCBI), peer-reviewed observational comparative study · 26 December 2024

   “In complex primary care cases, GPT-4 performs worse than human doctors taking the family medicine specialist examination.”

3. E3 Influence of a Large Language Model on Diagnostic Reasoning: A Randomized Clinical Vignette Study PubMed Central (NCBI), peer-reviewed randomized clinical vignette study · 14 March 2024

   “GPT-4 alone demonstrated higher performance than both physician groups, suggesting opportunities for further improvement in physician-AI collaboration in clinical practice.”

4. E4 Influence of a Large Language Model on Diagnostic Reasoning: A Randomized Clinical Vignette Study PubMed Central (NCBI), peer-reviewed randomized clinical vignette study · 14 March 2024

   “Our study and others were performed using clinical case vignettes that were curated and summarized by human clinicians with specific and answerable diagnostic questions in mind and do not reflect the full ambiguity in patient care settings.”

A 2025 systematic review and meta-analysis found no significant overall difference between generative AI models and physicians, and found that AI performed significantly worse than expert physicians (E1). A comparison on complex primary care cases found that GPT-4 performed worse than human doctors (E2). GPT-4 working alone did outscore physicians in one randomized study (E3), but that study used curated vignettes, and its authors say these do not reflect the full ambiguity of patient care settings (E4). In this evidence ChatGPT was ahead only on curated vignettes (E3). Elsewhere it showed no significant overall difference from physicians and performed worse than expert physicians (E1), or performed worse than doctors (E2). None of the evidence shows real-world results in which it broadly outperforms typical physicians, so the general claim is overstated.