Recently, the domestic large model DeepSeek V3 has drawn industry attention due to its outstanding performance in the AI arena. As the only open-source model to break into the top ten, it not only surpassed o1-mini but even exceeded Claude 3.5 Sonnet in various fields, including programming and mathematics. To verify its actual capabilities, a series of comparative tests were conducted. In the basic comprehension ability test, both models exhibited different characteristics. Faced with the Chinese riddle 'Xiaoming's mother has three children', DeepSeek V3 performed...