Google Gemini Unveils 'Thinking with 3 Pro' Mode for Interactive Scientific Exploration

Google DeepMind's Co-founder and Chief AGI Scientist, Shane Legg, recently announced the availability of a new "AI Mode" within the Google Gemini platform, dubbed "Thinking with 3 Pro." This advanced mode empowers users to create interactive applications for exploring complex scientific phenomena, marking a significant step in making sophisticated AI tools more accessible. Users can access this feature by navigating to the Gemini platform, selecting the AI Mode dropdown, and choosing "Thinking with 3 Pro."

The announcement, made via Legg's social media, highlighted the mode's potential for diverse applications, stating, "Now you can create interactive apps to explore all kinds of neat things! Doppler effect, orbital mechanics, black hole accretion disk, fluid dynamics..." This capability aligns with Google Gemini 3's enhanced reasoning and multimodal understanding, allowing it to generate code for high-fidelity visualizations and simulations. The "Thinking with 3 Pro" model is designed for "reasoning-heavy" tasks, distinguishing it from faster, less complex AI models.

This new functionality, internally codenamed "Nano Banana Pro" for its image generation capabilities, is built upon the recently released Gemini 3 Pro model. It offers breakthroughs in areas like text rendering, studio-level camera control, and character consistency, which contribute to the creation of detailed and dynamic interactive environments. These features enable users to transform abstract scientific concepts into tangible, explorable digital experiences.

The integration of "Thinking with 3 Pro" into the Gemini app, alongside its availability through Google Ads, Workspace, and the Gemini API, broadens its reach to a wide array of users, from researchers to educators and developers. This development underscores Google DeepMind's commitment to advancing AI's role in scientific discovery and interactive learning. The mode's ability to generate generative UI experiences and simulations directly from queries further enhances its utility in understanding complex topics.