Google Rolls Out Upgraded Gemini 2.0 AI for All Users

By siliconindia  |  Thursday, 06 February 2025, 10:26 IST

Google Rolls Out Upgraded Gemini 2.0 AI for All Users
Google has recently unveiled its highly developed AI model, Gemini 2.0, and its service is being opened to everybody with the fight between artificial and intelligent machines picking up pace in recent times. Google announced Gemini 2.0's wide release on 5 February when it made Gemini app's heavy-duty AI accessible both to computer users and people accessing it with their mobile handsets.
Although Google touts this as a 'general release', the company had initially provided access to select developers and trusted testers in December. Some features had already been integrated into Google’s products prior to this broader launch.
Last December, Google unveiled an experimental version of Gemini 2.0 Flash, which is an optimized model designed for developers with reduced latency and enhanced speed. Google later fine-tuned the Flash model by incorporating its rapid response capabilities with improved problem-solving abilities. This resulted in the updated Gemini 2.0 Flash Thinking Experiment released in Google AI Studio.
Building on this, with the February 5 update, Google has fine-tuned Gemini 2.0 Flash for even wider usage by users in both desktop and mobile environments. To that end, the company is also unveiling Gemini 2.0 Flash-Lite, the cheapest AI model from the company, and this is already open to public preview via Google AI Studio and Vertex AI.
The release is part of Google's larger strategy to invest heavily in AI-driven agents. The competition in AI technology has been increasing, especially with the rise of China's Deepseek. Industry giants such as Meta, Amazon, Microsoft, OpenAI, and Anthropic are also working on advanced AI models capable of executing complex multi-step tasks independently, reducing the need for continuous user guidance.
The latest AI models from Google support text output and multimodal input, with another set of additional features expected to be released in the coming months. For further details, including pricing, go to the Google for Developers blog.
According to Google, the Gemini 2.0 lineup has developed using novel reinforcement learning techniques; these methods let the model measure its responses with a self-review, which also improves its potential to handle rich prompts and thus deliver more relevant and focused answers.
As a means of building security, Google is using automatic red teaming to identify threats. This means countering indirect prompt injection attacks, where hackers inject malicious instructions into data an AI system is processing, in order to create a safer experience for all AI users.

ON THE DECK