Google has removed all restrictions for its Gemini 2.5 Pro artificial intelligence model, making it completely free for users with a Google account, ixbt.com reports. Previously, access to this version of the neural network was limited — now the technology, which leads the rankings in key metrics, has become publicly available.

Gemini 2.5 Pro outperforms competing models, including o1, o3-mini and Sonnet 3.7, especially in programming-related tasks. The system copes with both code generation and analysis, demonstrating higher accuracy and speed. The model also supports multimodality — it processes text, images, audio and video, combining data from different sources into a single context.

A special feature of Gemini 2.5 Pro is the ability to analyze up to 1 million tokens per request. Most modern analogues have a standard limit of 128-200 thousand tokens. This volume allows you to work with long documents, hours of video, or complex data sets while maintaining the integrity of the interpretation.