Google Expands Gemini 2.0 with Advanced AI Models

In Google AI Studio and Vertex AI, the company’s managed machine learning development platform, Google is first making a new version of Gemini 2.0 Flash widely accessible. This comes after the business made 2.0 Flash accessible to all customers through the desktop and mobile Gemini app. The business, Google, said on Wednesday that a number of its Gemini 2.0 artificial intelligence (AI) models were now available. The IT giant from Mountain View is introducing its Gemini 2.0 Flash Thinking Experimental AI model to web clients and mobile applications. In addition, an agentic version of the AI model that can communicate with specific apps is being released. Additionally, the business is giving premium subscribers access to an experimental version of Gemini 2.0 Pro. Furthermore, a public preview of a light version of the 2.0 Flash is also being released. The technology giant also said today that it is making more of its Gemini artificial intelligence models available and growing the family of models it already has.

The tech giant listed every model that is currently available to consumers in a blog post. Gemini’s free users can access some of these, while premium customers can access others, and developers are the only ones who can access others.

The Gemini 2.0 Flash Thinking, a reasoning-focused model similar to the DeepSeek-R1 and OpenAI’s o1 models, is the most prominent of them. Prior to its first release in December 2024, it was exclusively available through Google’s AI Studio.

The business is now making the model available to all users of the Gemini app and website. The model selector option at the top of the interface will provide access to the new AI model. Interestingly, it’s unclear if there will be any rate restrictions on the free tier for employing the Thinking model.

In addition, users will be able to access an agentic version of Flash Thinking 2.0 through the tech giant. Apps like Google Maps, YouTube, and Google Search may all be used with this device. Users should be able to ask Gemini to do certain activities on various applications thanks to this connection. The degree of its capability is presently unknown.

Additionally, Google revealed that 2.0 Flash Thinking Experimental is now widely accessible and launched an experimental version of Gemini 2.0 Pro, the company’s flagship model with the greatest performance for coding and sophisticated instructions. The new 2.0 Flash Thinking model is a compact, quick AI model that is tailored for reasoning and logic.

Google is releasing an experimental version of Gemini 2.0 Pro, the high-performance frontier model of the 2.0 series, for Gemini Advanced users. According to reports, the model performs well in activities involving coding and mathematics as well as decomposing complicated situations. With a context window of two million tokens, this is the most sophisticated model offered by the tech giant. Additionally, the model’s application programming interface (API) will have the capability to invoke tools like code execution and Google Search. Vertex AI and Google AI Studio will also offer it.

Additionally, Gemini 2.0 Flash-Lite, a brand-new AI model that is intended to be the company’s most economical model, to the public and this will now be available to developers. According to the business, it maintains speed and affordability while providing higher performance than the 1.5 Flash. It takes multimodal input and has a one million token context window. Additionally, Google is using the Gemini API to make the 2.0 Flash paradigm available to developers. At the moment, it can perform text-based activities; in the future, the business plans to add picture production and text-to-speech capabilities.

Google claimed to have gained insightful input on the capabilities of their AI models by distributing early, experimental versions of Gemini 2.0 to developers and experienced users. The business wants to carry on that trend with the launch of the experimental version of Gemini 2.0 Pro.

With a context window of 2 million tokens, the experimental Gemini 2.0 Pro model can process large documents and films, or around 1.5 million words. It can also run code and make calls to programs like Google Search.

The Gemini 2.0 Pro is the replacement for Google’s flagship Gemini 1.5 Pro model, which was introduced in February of last year.