Google has unveiled Gemini 2.5, its latest and most advanced AI model to date. The first release in this series is Gemini 2.5 Pro Experimental, which introduces built-in "thinking" capabilities. This new model is designed to tackle complex problems by breaking down tasks into multiple steps and reasoning through them before responding.
Gemini 2.5 Pro Experimental combines an enhanced base model with improved post-training, resulting in state-of-the-art performance across various benchmarks. It has achieved the top position on the LMArena leaderboard, which measures human preferences, by a significant margin. The model also excels in math and science benchmarks, outperforming competitors without relying on costly test-time techniques like majority voting.
One of the key features of Gemini 2.5 is its ability to analyze information, draw logical conclusions, incorporate context and nuance, and make informed decisions. This reasoning capability allows the model to handle more complex problems and support context-aware agents.
Gemini 2.5 Pro Experimental comes with a context window of 1 million tokens, enabling it to process extensive amounts of information, including large documents, audio, and videos. Google plans to expand this window to 2 million tokens in the future.
The model is now available for developers and enterprises through Google AI Studio, with plans to roll it out to Gemini Advanced users and Google Cloud's Vertex AI platform. For Samsung Galaxy S25 series users, there's a special offer of a free 6-month Gemini Advanced subscription, allowing them to experience Gemini 2.5 Pro at no cost during that period.
Google's introduction of Gemini 2.5 Pro Experimental marks a significant advancement in AI technology, positioning it as a strong competitor to other leading AI models in the market.
