Google has released its latest AI model “Gemini” to further its AI development and take on rivals like OpenAI GPT-4. The unique thing about Gemini is its multimodal approach. Multiple images, texts, code, audio, and video were used to train the AI. This allows the AI to generate results based on all these data.
Google Gemini Versions
Google’s Gemini AI model will have 3 models based on complexity and use.
- Gemini Nano
- Gemini Pro
- Gemini Ultra
The Gemini Nano is designed for smaller-scale on-device tasks. While the Gemini Pro serves as a versatile middle player with a scalable range of usage. Including its integration into the Google Bard Chatbot. While the Gemini Ultra is their Flagship version for handling the most complex tasks.
The use case scenarios are vast for these models and Google is going to deploy them accordingly. Gemini Nano will be powering the on-device services like reply suggestions in apps like WhatsApp. This will help in providing much better, contextual suggestions based on the content.
Google Gemini To Power Android
This Is just one of the many tasks Gemini Pro can accomplish. Additionally, Gemini Nano will power the future Android Smartphones and presently the Google Pixel 8 Pro. It’ll power the Android 14’s AI core and the Android developers can soon access it.
Moreover, Google’s Gemini Pro is scalable and can handle a wide variety of tasks like powering the AI chatbot Google Bard. We can test this new AI model in the Google Bard starting today. The service is available currently in more than 170 countries. However, it will work only in the English language for now. The inclusion of Gemini Pro will further the chatbot’s ability in reasoning, understanding and planning.
Google Gemini Ultra Coming Soon
As for the Gemini Ultra is concerned, Google is playing safe with it with vigorous testings. These tests are related to safety and trust is underway for the AI model. Only select developers and experts will get access to the Ultra model to receive feedback. Additionally, the special feature of the Ultra model is the integration of Reinforcement Learning From Human Feedback (called RLHF).
Moreover, Google Bard Advanced, a new Bard Chatbot version will be released in 2024. This will have advanced and latest Google AI models including the Gemini Ultra. Hassabis stated that Gemini, which is also enhancing Google’s generative AI search, will integrate into more of the company’s products and services, such as Search, Ads, Chrome, and Duet AI, in the coming months