A naukri.com initiative
TechCrunch
1w
102
Image Credit: TechCrunch
Google Gemini: Everything you need to know about the generative AI models
- Gemini is Google's suite of generative AI models, apps, and services, consisting of four different models: Gemini Ultra, Gemini Pro, Gemini Flash, and Gemini Nano.
- Gemini models are multimodal, meaning they can work with and analyze more than just text, with pre-training and fine-tuning on a variety of public, proprietary, and licensed audio, images, and videos, in different languages.
- The Gemini apps (formerly Bard) serve as front ends for Google’s generative AI, analogous to ChatGPT and Anthropic’s Claude family of apps.
- Gemini Advanced, accessible through the Google One AI Premium Plan, brings more sophisticated Gemini models to the Gemini apps with extras like priority access to new features, code execution, and a larger context window.
- Google's Gemini models' abilities span transcribing speech to captioning images and videos in real-time, while Imagen 3 model can generate creative images with accurate text prompt translations.
- Gemini models are available through Google's Gemini API for building apps and services, and pay-as-you-go, with the highest-priced Gemini 1.5 Flash costing 60 cents per one million output tokens.
- Google offered its own ethical warnings regarding the murkiness of training models on public data, and that certain users of Google Cloud can be shielded from lawsuits with the AI indemnification policy.
- Apple is reportedly in talks to utilize Gemini and other third-party models for the Apple Intelligence suite.
- As of September 2021, Gemini Ultra and 2.0 Flash pricing has yet to be announced, and Nano is still in early access, with Gemini lacking fixes for issues with generative AI like biases and hallucinations.
- Google has shown off its multipurpose model with Project Astra, designed for real-time, multimodal understanding, with hopes to use it in smart glasses.
Read Full Article
6 Likes
For uninterrupted reading, download the app