TMTPOST – Google has launched its next-generation Gemini 2.5 series of AI reasoning models, introducing a breakthrough in long-context understanding with support for up to 1 million tokens—equivalent to approximately 750,000 English words—in a single prompt. The first released model, Gemini 2.5 Pro Experimental, debuted Tuesday on Google AI Studio and will be available to subscribers of the $20/month Gemini Advanced service. Google claims this multimodal reasoning model is currently its "most intelligent" AI system, featuring a "deliberative" architecture that pauses to "think" before generating responses.