Translated data: Gemini Pro demonstrates significant potential in common sense reasoning, challenging previous evaluation results. On par with GPT-3.5, Gemini Pro slightly outperforms in specific tasks. Reasoning experiments show that both Gemini Pro and GPT-4Turbo exhibit advanced reasoning mechanisms in both correct and incorrect answers.