Google just flipped a switch inside the Gemini app that changes how it explains hard stuff. Instead of dumping a wall of text or a static diagram, Gemini can now spin up custom, interactive simulations and 3D models right in your chat—things you can rotate, tweak, and explore as if you were playing with a mini‑app built on the fly.

The company announced the feature in a Google Workspace Updates post dated August 23, 2026, saying Gemini can “transform your questions and complex topics into custom, interactive visualizations — directly within your Gemini app chat.” It’s a notable step up from earlier behavior, where responses were “largely text with static diagrams,” according to Google’s own description. Now, the app delivers “functional simulations” designed to help you actually understand the topic you asked about—whether that’s rotating a molecule, probing a physics system, or visualizing a process that’s tough to picture in your head.

Imagine asking Gemini to “show me how a pendulum’s period changes with length” or “help me visualize this protein’s structure.” With the new capability, the reply can include sliders, controls, and a live 3D view you manipulate directly in the conversation. You’re not just reading about velocity or gravity—you can adjust those variables and watch the simulation respond in real time.

Google frames this as moving from explanation to exploration. The responses can include “visual elements — like tables, grids, and simulations — made specifically for your question,” the company writes, emphasizing that each visualization is tailored rather than pulled from a fixed library. In practical terms, that means two people asking related but different questions could get two different interactive models, each tuned to the specifics of their prompt.

Although the latest announcement landed in late August 2026, the underlying capability has been in motion for months. Google noted in April 2026 that the Gemini app “can now generate interactive simulations and models,” and that the feature was rolling out globally to all Gemini app users. By the time of the August post, the functionality was described as available now to all Google Workspace customers, Workspace Individual subscribers, and eligible users, with admins seeing it turned on by default for organizations that have Gemini enabled.

This sits inside a bigger wave of 2026 updates aimed at making Gemini more visual, more agentic, and more integrated into daily workflows. Earlier this year, Google introduced a redesigned “Neural Expressive” interface that favors dynamic graphics, interactive timelines, and rich media over plain text blocks. It also rolled out faster models like Gemini 3.5 Flash for complex tasks, and agent-style helpers such as Daily Brief and Gemini Spark that proactively organize information and automate routines. The simulation feature is the most obviously “hands-on” piece of that puzzle: instead of just summarizing or planning, Gemini can now build you a little interactive lab inside the chat.

For students and educators, the shift from static to interactive is a big deal. Google has been positioning Gemini for education as a tool to “create personalized learning experiences” and help learners “learn confidently,” and interactive models fit neatly into that promise. Rather than staring at a textbook diagram of a molecule or a circuit, a student can twist it in 3D, isolate parts, and test “what if” scenarios on the spot. Google’s own student-facing messaging highlights “interactive visualizations” as a core study aid, alongside features like study notebooks and Gemini Live.

For professionals and curious adults, the value is similar: complex systems become easier to grasp when you can play with them. Engineers, analysts, and product folks can ask Gemini to “help me visualize” a workflow, a data relationship, or a mechanical concept and get something closer to a prototype than a paragraph. That’s especially useful in a world where AI is increasingly expected not just to answer questions but to help you think through them.

There are limits, of course. Google notes that “usage limits apply,” and the feature is governed by the same Generative AI settings that control other Gemini tools in Workspace. End users don’t get a separate on/off switch; if your organization has Gemini enabled, this capability is on by default. Still, even with constraints, the direction is clear: Google wants Gemini’s explanations to feel less like reading a manual and more like tinkering with a model on your desk.

How to try it right now

If you have access to the Gemini app and the Pro model, you can start experimenting immediately. Google’s guidance is straightforward: head to gemini.google.com, select the Pro model in the prompt bar, and ask Gemini to “show me” or “help me visualize” a complex concept. From there, you can nudge it toward specific kinds of simulations—physics demos, molecular structures, system diagrams—and then use the on-screen controls to explore.

In a year where AI assistants are racing to become more agentic and more integrated, this feature stands out for being tactile. It doesn’t just tell you how something works; it hands you the knobs and lets you find out for yourself.