Google’s Gemini app just got a big upgrade: it can now answer your questions with interactive simulations, 3D-style models, and live-updating charts right inside the chat window. Instead of only dumping long text or static diagrams, Gemini can literally show you what’s happening and let you play with the variables.

Think of asking, “How does the moon orbit the Earth?” and, instead of a wall of explanation, getting a live simulation where you drag sliders for gravity and initial velocity and instantly watch the orbit become stable, spiral out, or crash. You’re not locked into one canned animation; you can tweak the setup in real time to actually see why the physics behaves a certain way.

It’s not limited to space stuff, either. You can ask Gemini to “show me how fractals work,” “help me visualize supply and demand,” or “plot how my savings grow at different interest rates.” Gemini then turns that into custom interactive visuals — from rotating molecules to dynamic charts that update as soon as you change a number. Instead of hopping to a separate graphing tool or coding environment, everything runs directly inside the chat window.

Using it is intentionally low-friction: go to gemini.google.com or the Gemini app, switch to the Pro model, and use prompts like “show me…” or “help me visualize…” followed by the concept you’re trying to understand. Gemini then decides whether a 3D model, physics simulation, diagram, or chart makes the most sense and builds that on the fly, with controls you can poke at — sliders, toggles, number fields, zoom, rotation, and pause/play for time-based simulations.

Under the hood, Google is essentially rendering web-style interactive experiences (think WebGL + JavaScript) right inside the chat UI, sandboxes and all, so the AI can safely run these little “mini apps” without opening a new tab or sending you elsewhere. There is a tradeoff: Workspace and Education accounts don’t get this yet, which Google is reportedly tying to the high compute cost of running rich, interactive visualizations at scale.

This move also quietly raises the bar for AI chatbots. Anthropic’s Claude recently started auto-generating charts and interactive diagrams, and OpenAI has been pushing more visual reasoning tools in ChatGPT. Google’s answer is clear: if you’re going to spend time chatting with an AI, it should feel more like a lab or a sandbox than a static answer box.

For learners, creators, and even casual users, the practical impact is simple: instead of passively reading explanations, you can now experiment. Nudge a slider, change a constant, rotate a structure, re-run a scenario — and watch the concept reshape itself in front of you. It’s the difference between being told how something works and getting to poke at the system until it finally clicks.