Gemini is Google’s long-promised, next-gen generative AI model family. © 2024 TechCrunch. All rights reserved. For personal use only.
Google's Gemini is a next-generation generative AI model family developed by DeepMind and Google Research. It includes variants like Gemini Ultra (very large), Gemini Pro (large), Gemini Flash (faster, "distilled" version), and Gemini Nano (small, offline-capable models). Gemini is natively multimodal, meaning it can process and generate text, images, audio, and video, setting it apart from earlier models like Google's LaMDA, which was text-only.
Gemini models are accessible via Gemini apps, which serve as chatbot interfaces for interacting with the AI. These apps are available on the web, Android (replacing the Google Assistant app), and iOS (via Google and Google Search apps). The Gemini apps support images, voice commands, text, and soon, videos, and they integrate with Google Drive.
Gemini Advanced, available through the Google One AI Premium Plan ($20/month), offers enhanced features, including access to more sophisticated Gemini models, priority feature access, extended context windows (750,000 words vs. 24,000 words for standard Gemini), and tools like Deep Research and NotebookLM. Ethical concerns include the use of public data for training, with unclear legal implications, though Google offers limited indemnification for certain Google Cloud customers.
