斯坦福临床医疗AI横评，DeepSeek把谷歌OpenAI都秒了

Stanford's latest evaluation of clinical AI models ranks DeepSeek R1 first with a 66% win rate, outperforming Google and OpenAI. This assessment focuses on real-world scenarios faced by clinicians, highlighting DeepSeek's practical application in healthcare.

Key Points

• DeepSeek R1 achieved a 66% win rate in the Stanford evaluation.
• The evaluation emphasizes real-world clinical scenarios for AI models.
• Google and OpenAI were outperformed by DeepSeek in this benchmark.
• The results have impressed international audiences, particularly in the US.
• This assessment could influence future AI applications in healthcare.