Google Launches Gemini 2.0 AI Model Powered by Trillium Chip, Offering 2x Speed and Enhanced Multimodal Features

Dec 11, 2024, 03:37 PM

Google has announced the release of Gemini 2.0, its latest and most capable AI model designed for the agentic era. The new model, Gemini 2.0 Flash Experimental, outperforms the previous 1.5 Pro model at twice the speed. It offers enhanced multimodal capabilities, including text, image, and speech generation, as well as native image editing and support for voice-to-voice interactions. Gemini 2.0 features a 128k tokens context window and introduces new native tool use and a Multimodal Live API. Google also unveiled its sixth-generation Trillium AI chip, which delivers four times the training performance and powers Gemini 2.0. Developers can access an experimental version of Gemini 2.0 Flash through Google AI Studio and Vertex AI. The model is designed to power AI agents, including Project Astra and agents for gaming and web browsing, paving the way for new agentic experiences.