OpenAI continues to be a leading company in the artificial intelligence market thanks to the successful GPT-4. However, this seems to be about to change. And Mountain View has just launched Gemini, Google’s new language model that’s much more advanced than GPT-4 and has all the AI currently available. Next, we will tell you why this is the most powerful artificial intelligence and how to test it.
Google Gemini: Google’s new AI based on a multimodal model that is more powerful than GPT-4
As Google explained on its blog, Gemini is a new multimodal language model that is the most powerful artificial intelligence to date. And if you want to check it out, We strongly invite you to watch the video. We’ll leave it here to show off all that the Google team can do.
Well, this AI has the ability to understand and generate written, visual, auditory and coded information. Hence its “multimodal” nature. This gives it superior capabilities to most language models (LLM). In fact, the AIs outperformed OpenAI’s GPT-4 by scoring the best in 30 of the 32 academic tests tested.
As many as 57 subjects to assess knowledge and problem-solving skills in areas such as mathematics, physics, history, law, medicine and ethics, MMLU scores 90%, surpassing even human intelligence. .
What else makes it different? Well, Google Gemini shines in the way it interacts without the need for text. This makes for a more fluid experience for the user, as they can interact with the AI through conversation or visual contact.
How to use Google Gemini
Google is launching Gemini in 3 versions Nano, Pro and Ultra. The most powerful version is Gemini Ultra and will arrive in early 2024. Nano is very powerful, but it is fast and made for simple tasks from the device.
Meanwhile, Gemini Pro is now available. Google Bard has been updated and adopted this multimodal language model, however Currently only in English. So, all you have to do to try Gemini is to use Google Bard. We leave the link below:
To try Google Gemini | Google Bard
And you… do you think Google will be the benchmark company in AI, or do you think OpenAI will have something up its sleeve?
