Google Unleashes Its Most Powerful AI Models for All, Advancing Virtual Agents

Google unveiled Gemini 2.0, its most advanced AI model suite, making it accessible to everyone. Previously available only to developers and select testers since December, the release now marks a general launch.

The Gemini 2.0 suite includes three models: Flash 2.0, optimized for handling large-scale, high-frequency tasks; Pro Experimental, designed for advanced coding tasks; and Flash-Lite, the company's most cost-efficient model. Pricing starts at $0.10 per million tokens for Flash, while Flash-Lite costs $0.0075 per million tokens for text, image, and video inputs.

This release is part of Google's push to lead in AI agent technology, where models can independently complete complex tasks. Rival companies like Meta, Amazon, Microsoft, OpenAI, and Anthropic are also advancing in this area. These models aim to function as sophisticated assistants capable of planning, analyzing, and making decisions on behalf of users.

Anthropic has developed AI agents that operate computers similarly to humans, navigating websites and executing complex tasks. OpenAI's "Operator" feature helps automate tasks like booking reservations and managing forms. Meanwhile, Google's Deep Research, launched in December, assists in exploring complex topics and compiling research reports.

CEO Sundar Pichai emphasized the importance of executing well in this competitive landscape, targeting 2025 as a pivotal year for innovation.

Source: cnbc.com