Google has unveiled its upgraded AI model, Gemini 3, which features significant enhancements to performance and functionality. Released on the 18th, Gemini 3 is described as a multimodal model capable of processing text, images, audio, and video in a unified flow. Notably, Google claims it ranks highly on various AI benchmark tests, demonstrating superior logic, mathematics, and fact-checking abilities. The model now provides more concise and direct responses while reducing bias compared to its predecessors.
Further advancements include the ability to generate presentation and webpage layouts based on user prompts, facilitated by an enhanced query fan-out search feature for better understanding of user intent. The AI mode in the Gemini app and Google Search has also been revamped to deliver answers incorporating images, tables, grids, and interactive simulations. Additionally, Gemini 3 Pro boasts improved reasoning and planning capabilities, allowing it to handle complex, long-term tasks. It operates using Gemini Agent, Google’s experimental tool, which automates tasks like email organization and travel planning. Currently, Gemini 3 Pro is available in preview within the Gemini app for subscription users in the U.S. alongside advanced AI functionalities.
👉 Pročitaj original: CIO Magazine
