According to Oriol Vinyals, the updated and faster Gemini 3 Deep Think has been unveiled, showcasing PhD-level reasoning capabilities for tackling rigorous STEM challenges. This AI model delivers high-performance results, including gold medal-level achievements in Physics and Chemistry Olympiads, making it a significant advancement in AI-driven problem solving.

According to Demis Hassabis, Gemini 2.5 Pro is the leading AI model on LMArena, surpassing competitors by 39 ELO points. This advancement is due to significant improvements in multimodal reasoning, coding, and STEM capabilities, which could influence AI-driven trading algorithms and enhance data processing efficiency. These updates make it a valuable tool in AI Studio and GeminiApp for developing sophisticated trading models.