Overview

• Gemini 2.5 Pro Experimental debuts with advanced reasoning capabilities, achieving top scores on benchmarks like Humanity's Last Exam with 18.8%.
• The model supports multimodal inputs, including text, audio, images, video, and code, with a 1 million token context window, soon expanding to 2 million.
• Available now to developers via Google AI Studio and Gemini Advanced subscribers, with Vertex AI support expected in the coming weeks.
• Gemini 2.5 Pro outperforms competitors in coding benchmarks such as Aider Polyglot and SWE-bench Verified, demonstrating strong coding and problem-solving abilities.
• Google plans to integrate reasoning capabilities into all future AI models, advancing its strategy for more context-aware and capable systems.