Google has unveiled Gemini 2.5, its latest family of AI reasoning models.
The Gemini 2.5 models are designed as "thinking models," capable of analyzing information, drawing logical conclusions, and incorporating context and nuance before responding.
“Gemini 2.5 is a thinking model, designed to tackle increasingly complex problems,” Google Deepmind chief technology officer Koray Kavukcuoglu said in a blog post.
“Our first 2.5 model, Gemini 2.5 Pro Experimental, leads common benchmarks by meaningful margins and showcases strong reasoning and code capabilities.”
The initial release includes Gemini 2.5 Pro Experimental, which Google describes as its most intelligent model yet.
According to the company, Gemini 2.5 Pro Experimental outperforms previous models and some competing AI models on several benchmarks.
This includes leading in math and science benchmarks, scoring 18.8% on Humanity's Last Exam, a dataset designed to assess AI's ability to handle complex knowledge-based questions.
It also demonstrates strong performance on common coding, math, and science benchmarks like GPQA and AIME 2025.
This model is now available on Google AI Studio and in the Gemini app for subscribers to the company's $20-a-month AI plan, Gemini Advanced.