Google has a metal pedal to develop artificial intelligence. Just a few months after its appearance Gemini 2.0The technology giant revealed another upgrade in Gemini 2.5. As with any new launch of Amnesty International, Google promotes a powerful performance on the LMARNA of GEMINI 2.5, especially its capabilities in coding, mathematics and science.
The first model in this series is Gemini 2.5 Pro experimental. This is a thinking model that aims to provide responses based on more thinking, analysis and context more than the answers provided by models that depend on classification and classification. It is a different approach from the Google followed by the Gemini 2.0 series, which started with a more efficient and lowest flash release.
"With Gemini 2.5, we have achieved a new level of performance by combining a largely enhanced basic model with the improvement of training," The company said in a blog post attributed to Korai Kavokooglu, CTO from Google DeepMind. "To move forward, we build these thinking possibilities directly in all our models, so that they can deal with more complex problems and support more capable agents."
Google has just started offering Gemini 2.0 to its services, using it to operate the addition recently Artificial intelligence mode In the search and Deep search To deal with more complicated inquiries. With today’s launch, he expected to hear more updates from the company about getting this latest version. Gemini 2.5 Pro experimental is now available in Google Ai Studio, and Gemini Advanced members can use it directly in the Gemini app.
This article was originally appeared on Engadget on https://www.engadget.com/ai/google-release-gemini-25-ai-model-for-complex-thinking-18235224.html?
[og_img]
Source link