Google is opening a new era in the field of artificial intelligence with a series of updates to the Gemini model family. The first local multimodal model, Gemini 1.0, offered in three different sizes (Ultra, Pro and Nano) in December, soon received the 1.5 Pro version with improved performance and an expanded context window of 1 million tokens.
In light of user feedback, Google introduced Gemini 1.5 Flash to meet the need for lower latency and service cost. Lighter than the 1.5 Pro, this model is optimized for speed and efficiency and is ideal for high-volume, high-frequency tasks. Offering an expanded context window of 1 million tokens, 1.5 Flash shows superior performance in tasks such as summarization, chat applications, image and video subtitling, and data extraction from long documents and tables.
Google has also significantly improved the 1.5 Pro, which is the best model for overall performance. Expanded to 2 million tokens, the context window delivers improved performance in features such as code generation, logical reasoning and planning, multi-round conversation, and audio and video understanding, powered by data and algorithmic optimizations. 1.5 Pro can now follow more complex and nuanced instructions, including product-level behavioral determinants such as role, format, and style.
Gemini Nano, on the other hand, can go beyond text input and process images as a network. Starting with Pixel phones, apps that use Gemini Nano with Multimodality will be able to understand the world the same way humans do.
In order to benefit humanity, Google DeepMind is moving towards developing universal artificial intelligence agents that can help in daily life with Project Astra. Astra aims to develop AI agents that can understand context and take action the same way humans understand and react to a complex world. Designed as proactive, approachable and personalized assistants, these agents will be able to interact with users naturally and without delay.
Astra was designed to be able to process and remember video and speech input. Agents built on the Gemini model and other task-specific models process information faster by continuously encoding video frames, combining video and speech input into a timeline of events, and caching that information for efficient recall.
Google also continues to develop its Gemma family of open models. Gemma 2, the next generation of open models for responsible AI innovation, will feature a new architecture for breakthrough performance and efficiency and will be available in new sizes.
Google is constantly improving the Gemini model family to shape the future of artificial intelligence, providing users with access to smarter and more useful tools in their daily lives.
Google Gemini is being personalized!
Google continues to develop its personal artificial intelligence assistant Gemini. Designed as a conversational, intuitive and helpful assistant, Gemini helps you tackle complex tasks and take action on your behalf. Gemini, which you can use in the app or via the web experience, is constantly updated.
Analyze documents with the world’s longest context window
Google offers its newest model, Gemini 1.5 Pro, to Gemini Advanced subscribers. Gemini 1.5 Pro, the consumer chatbot with the world’s longest context window, offers an extended context window starting from 1 million tokens. This means Gemini Advanced can make sense of multiple large documents totaling up to 1,500 pages or summarize 100 emails. Soon, it will be able to process an hour of video content or more than 30,000 lines of code base.
To take advantage of this expanded context window, you can upload your files to Gemini Advanced via Google Drive or directly from your device. Now you can quickly get answers and information on dense documents, like finding the details of the pet policy in your lease or comparing the key arguments of multiple lengthy research papers. Soon, Gemini Advanced will act like a data analyst, surfacing insights and instantly creating custom visualizations and charts from uploaded data files like spreadsheets.
More natural conversations with Gemini Live
Google is introducing new ways to interact with Gemini more naturally. With Gemini in Google Messages, you can now chat with Gemini in the same app where you message your friends.
In the coming months, a new mobile chat experience, Live, will be available for Gemini Advanced subscribers. This feature makes talking to Gemini more intuitive, using Google’s most advanced speech technology. With Gemini Live, you can talk to Gemini and choose from a variety of natural voices for it to respond. Just like any conversation, you can speak at your own pace or interrupt mid-response by asking clarifying questions.
Creating complex plans just got easier
Travel planning often takes more time than the trip itself. Gemini Advanced’s new planning experience goes beyond showing you a list of suggested activities and creates a custom itinerary for you.
For example, tell Gemini, “My family and I are going to Miami for Labor Day. My son likes art and my husband wants fresh seafood. Can you get my flight and hotel information from Gmail and help me plan the weekend?” you may ask.
This request requires Gemini to do much more than just present publicly available information like other chatbots. Gemini takes into account your flight timing, dining preferences, and information about local museums, while also understanding where each stop is and how long it will take to travel between each activity. It pulls your flight information from Gmail, taps Google Maps for restaurant and museum recommendations near your hotel, and uses Search to suggest other activities to fill the rest of your day, like a walking tour in the Design District or beach time. It synthesizes all this information for you and creates a personal, customized itinerary that meets all your wishes. If you make changes or add more details, the itinerary is automatically updated.
Personalize Gemini with Gems
For an even more personalized experience, Gemini Advanced subscribers will soon be able to create Gems. Gem is a customized version of Gemini. You can create any Gem you can imagine: a gym buddy, sous chef, coding partner, or creative writing mentor.
Installing Gem is also very easy. Simply describe what you want your Gem to do and how you want him to respond (e.g., “You are my running coach, give me a daily running plan and be positive, cheerful, and motivating”). Gemini will take these instructions and refine them with a single click, creating a Gem that meets your wishes.
Connect with more Google apps
Last year, Google brought Extensions directly to Gemini, allowing you to do more with the Google apps and services you already use. Currently, Google applications such as YouTube Music Extension are being integrated into Gemini.
More Google tools like Google Calendar, Tasks, and Keep will soon be connected to Gemini. This way, you can take a photo of your child’s school curriculum and ask Gemini to create a calendar entry for each assignment.
With these updates, Google ensures that Gemini becomes a personal and customizable artificial intelligence assistant that can better respond to users’ needs.