As usual, Google affirms the outstanding performance of Gemini 2.5 especially in the fields of programming, toancture and science.
The first version in the upgraded lineup is the Gemini 2.5 Pro Experimental. According to Google, this is a model designed to think - providing feedback based on argumentation, analysis and context, rather than simply classifying or predicting like traditional artificial intelligence (AI) models.
This is also a different approach from the previous Gemini 2.0 line, which started with the Flash version - optimized for speed and cost.
With Gemini 2.5, we have reached a new level of performance thanks to the combination of a strongly improved background model and post-upschooling. From now on, we will integrate this thinking ability directly into all models, helping them solve more complex problems and supporting smarter tasks with context recognition, said Koray Kavukcuoglu, technology director at Google DeepMind, in a blog post.
Previously, Google had only just begun rolling out Gemini 2.0 to its services, including AI Mode in the search engine and the Deep Research feature for in-depth queries.
With the launch of Gemini 2.5, users can expect to receive more updates from Google about the integration of this latest version soon.
Gemini 2.5 Pro Experimental is now available on Google AI Studio and Gemini Advanced package users can experience it directly through the Gemini application.
