Gemini 2.5 Pro is our most advanced reasoning Gemini model, capable of solving complex problems. It features a 2M token context window and supports multimodal inputs including text, images, audio, video, and PDF documents.