Google DeepMind launches its Gemini 2.5 Deep Think AI model, touted as its most advanced reasoning system. Available to $250 monthly subscribers, the model uniquely tests multiple ideas in parallel, generating superior answers. It demonstrated a 34.8% score on the rigorous Humanity’s Last Exam, outperforming rivals like xAI's Grok 4 and OpenAI's o3. This innovative AI, which can take hours to reason, aims to enhance research and creative problem-solving capabilities in various fields.