Google launches Gemini I have improved with a new “thought” model

Google has unveiled improvements to Gemini, its artificial intelligence product designed to compete with the Chatgpt, Deepseek and Apple Intelligence of Openai.

Users of the Gemini application can now use the Google Flash Thinking 2.0 Experimental model, which is specifically designed to deconstruct prompts in a sequence of steps, thus improving its reasoning capacities.

Gemini Live arrives in more than 40 languages! It takes place on Android devices in the coming weeks, starting with French, German, Portuguese, Hindi and Spanish, with many others on the way.

Soon you can communicate, collaborate and be creative even more … pic.twitter.com/yp4yvfwfcr
– Google Gemini application (@GeMinIApp) October 3, 2024

This updated model transparently displays its reasoning process, providing users with a clearer understanding of their responses.

In addition, a version of Flash 2.0 thought is able to interact with applications such as YouTube, Search and Google Maps, because Google explores how these new reasoning features can help users maximize their experience with Google applications.

An experimental variant known as Gemini 2.0 Pro has also been introduced, which, according to Google, offers higher performance in coding and the fight against complex prompts.

Gemini 2.0 Pro is accessible via Google AI Studio, Vertex AI and the Gemini application for Gemini Advanced subscribers. In addition, a Gemini 2.0 Flash-Lite model is available in Google AI Studio, presented by Google as the most profitable model to date.

https://www.youtube.com/watch?v=MLBGFHTMZYK

Experimental models Gemini 2.0 Flash Thinking and 2.0 Pro are gradually deployed on Gemini web and mobile applications, where they can be selected from the drop -down menu during AI interactions.

Source link