Translated data: Google has introduced the Gemini AI model, claiming superiority over OpenAI's GPT-4 in academic benchmark tests. The slight advantage of Gemini leaves users hesitant about switching to Bard. However, the points of comparison are becoming increasingly abstract and nearly meaningless. In practical applications, Gemini still has errors. Google needs to work harder to achieve a passing score.