Google has introduced Gemini 2.5 Pro, its latest and most advanced AI model, designed for complex problem-solving. Available now for developers in Google AI Studio and Gemini Advanced subscribers, this enhanced model boasts a “thinking” approach to processing information. Vertex AI support is expected soon.
The Gemini Advanced logo
Google claims Gemini 2.5 Pro outperforms competitors, securing the top spot on the LMArena leaderboard with an 18.8% score. This surpasses other leading AI models like ChatGPT and Deepseek. This achievement highlights the model’s superior performance in advanced reasoning tasks. Specifically, Gemini 2.5 Pro excels in math and science benchmarks such as GPQA and AIME 2025 without relying on costly test-time techniques like majority voting. It also achieves a state-of-the-art score on Humanity’s Last Exam, a dataset designed to test the limits of human knowledge and reasoning.
Gemini Advanced subscribers can access 2.5 Pro at no additional cost. For larger-scale usage, pricing details will be released in the coming weeks. The model currently offers a one-million context window, with a two-million window expansion planned for the near future. Furthermore, Gemini 2.5 Pro can process complex information from diverse sources, including audio, text, images, code repositories, and video.
Since the launch of the “thinking” version in December, Google has continued to expand Gemini’s capabilities. Recent integrations include Google Maps, enabling users to interact with the platform conversationally. Google also announced plans to replace Google Assistant with Gemini in the coming months.
This continuous development underscores Google’s commitment to advancing AI technology and integrating it into its core products and services.
