TMTPOST – Google has launched its most advanced AI model yet, the Gemini 2.5 series, with the experimental Gemini 2.5 Pro as the first inference model in the lineup. Gemini 2.5 Pro is designed to tackle complex tasks with exceptional reasoning and coding capabilities, incorporating all features from previous Gemini models. In benchmark tests like GPQA and AIME 2025, Gemini 2.5 Pro has outperformed competitors, securing the top spot in the Chatbot Arena with a 39-point lead over Grok-3 and GPT-4.5, demonstrating its superior performance in human-preference tests.