Google Unveils Gemini 2.0 Flash and AI Agents

Gemini 2.0 Flash: Power and Speed

Google has announced the new Gemini 2.0 Flash model, surpassing its predecessor, Gemini 1.5 Pro, in performance, speed, and multimodal capabilities. The model is twice as fast, supports multimodal output, including image generation, audio with text, and text-to-speech conversion. It can process input from images, videos, and audio, and can also invoke external tools like Google Search or execute code.

For developers, Gemini 2.0 Flash is available in AI Studio and Vertex AI, alongside the new Multimodal Live API, which supports real-time video and audio streaming. The model is set to debut in the Gemini user app in January 2025.

Innovative AI Agents from Google

In addition to Gemini 2.0 Flash, Google introduced projects with advanced agent capabilities:

• Project Astra offers multilingual communication support, including mixed languages, along with integration with Google Search, Lens, and Maps. 
• Project Marinerexcels in analyzing and interpreting browser data, demonstrating 83.5% efficiency.
• Jules, a coding development tool integrated with GitHub, simplifies workflows for programmers, enabling task planning and problem-solving under user guidance.
• Google’s new Deep Research tool leverages Gemini for online data retrieval and generating comprehensive analytical reports.