On December 6th local time, American technology company Google launched the Gemini model. Unlike the big language model ChatGPT, the Gemini big model is a native multimodal big model that can not only process text content, but also seamlessly process information from multiple modalities such as code, audio, images, and videos. According to internal personnel, Gemini has trillions of parameters and the computing power used for training can even reach five times that of GPT-4. Before releasing Gemini, Google conducted a series of tests to evaluate it using standard industry indicators. The results showed that Gemini Ultra surpassed GPT-4 in 30 out of 32 benchmark indicators, from natural image, audio, video understanding to mathematical reasoning. In the MMLU (Massive Multi Task Language Understanding) test, Gemini Ultra surpassed human experts for the first time with a high score of 90.0%. A year ago, ChatGPT launched by OpenAI went viral overnight, and a year later, Google launched Gemini to counter OpenAI and demonstrated comprehensive surpassing of GPT-4 in many capabilities. Who will laugh the last in the future, we will wait and see.