Fun games and (video) games using Google’s Gemini 3 AI model

AI Video & Visuals


Business Insider’s amazing Google reporter Hugh Langley has been playing with (or… working) with Google’s new Gemini 3 AI service this week.

This is the latest big AI model release and competes with OpenAI’s GPT-5, xAI’s Grok 4, and Anthropic’s latest offering.

Gemini 3 has gotten good reviews so far. Google stock hit an all-time high on Wednesday, putting the company’s market value very close to Microsoft’s.

Hugh’s first thoughts after playing Gemini 3 for a day or so:

“I think what’s most impressive about Gemini 3, and what’s already garnering attention, is its ability to create something new out of the whole, thanks to its improved multimodal and coding capabilities.”

Gemini 3 is particularly good at designing interactive simulations. Hugh told me it could be interesting for visual learning, building websites and apps, or just having fun.

He tried out this new model in a Google sandbox called AI Studio. This is similar to a real sandbox, but for developers. It is also digital and allows access to Google’s AI services over the internet. No real sand.

Hugh started with something basic: an interactive website about elephants.

“I asked Gemini to include lots of fun widgets and trivia about the animal, but not much else. I wanted to see how well Gemini would fill in the gaps.”

The overall design of the website was a little sparse, but it worked. Hugh loves this little widget that generates fun elephant information every time you press a button.


Gemini 3 screenshot

Gemini 3 screenshot

Hugh Langley/Google Gemini 3



“It also included a mini-game where you feed the elephants peanuts,” Hugh said. “When I filled the bar with water, a message popped up to let me know that the elephant was happy. That was nice.”


Gemini 3 screenshot

Gemini 3 screenshot

Hugh Langley/Google Gemini 3



Being able to visualize complex ideas is an area where AI can be particularly useful. Hugh asked Gemini 3 to create an interactive website to explain photosynthesis.

“Gemini has produced several sliders to adjust the levels of sunlight, water and carbon dioxide that float around it as particles of different colors,” he said. “If the balance is correct, we have successfully generated energy!”

Hugh wasn’t producing energy. In fact, running the Gemini 3 model probably consumed a lot of power. Still, it’s pretty impressive.


Gemini 3 screenshot

Gemini 3 screenshot

Hugh Langley/Google Gemini 3



Other Gemini 3 users are also creating interesting new things. Hugh liked this idea of ​​having Gemini build a LEGO Creator through simple prompts such as “Build a 3D LEGO builder. Let me choose bricks of different shapes.”

Here is an example. Hugh was able to reproduce his version.

There are also all kinds of video games that users can create using Gemini 3. Jeff Dean, one of Google’s top AI researchers, posted some clips of these games.

Hugh also tried this. It took a little more interaction with Gemini to get this right.

He wanted to create “Super Dario Land,” a game where players had to guide Anthropic CEO Dario Amodei to the correct warp pipe. If successful, the player is awarded AGI (a moment when a machine theoretically outperforms a human on most tasks).


Google Gemini 3 screenshot

Google Gemini 3 screenshot

Hugh Langley/Google Gemini 3



Hugh asked Gemini to style it like an old Nintendo Game Boy game.

“At first, Dario couldn’t jump high enough, so he asked Gemini to fix it,” Hugh told me. “Just one more prompt and physics is solved.”

Gemini worked to map the controls to Hugh’s computer keyboard without any instructions from Hugh, so the game was playable very quickly.

“The game itself may not be a hit, but I’ll leave the door open for a sequel. Dario, give me a call!” Hugh said.

If you’d like to play Hugh’s game, email him at hlangley@businessinsider.com. It’s fun!

Sign up for BI’s Tech Memo newsletter here. Please contact us by email. abarr@businessinsider.com.





Source link