We built a live, player-generated game at #GoogleIO using the Gemini API, Nano Banana, @GoogleAIStudio, @Antigravity and Lyria 3. 🕹️

Here’s how we built Infinite Scaler using our AI tools ↓ 1️⃣ First, we used the Gemini API to take a simple prompt like: “a majestic space fox levitating in space with a giant saturn and stars behind” and expand it.

This mapped out all the colors, backgrounds, and decorations needed to build a level in the game.

8 surprising ways we used AI at #GoogleIO this year 🧵↓ 1️⃣ Branding: This year’s visual brand identity was a close collaboration between our team and AI.

We started by giving Gemini previous brand guidelines and five years of I/O recaps.

We then generated new imagery and iteratively fed outputs back into Nano Banana with feedback. We also used Nano Banana to explore icon styles.

Here’s a prompt we used to explore icon styles with Nano Banana in the @GeminiApp:

You are an expert image editor. You will be given two images.

**Image 1** provides the **texture and material**.

**Image 2** provides the **pose, shape, and lighting**.

Your task is to create a new image by applying the detailed texture and pattern from the icon in **Image 1** onto the surface of the white icon in **Image 2**.

**Crucial Constraint:** The output image must perfectly preserve the exact pose, camera angle, scale, and lighting of the icon in **Image 2**. Do not change its orientation or position in any way. The final result should be the icon from Image 2, but with the texture of Image 1.

Meet DiffusionGemma ⚡ Our latest experimental open model (Apache 2.0) that generates text up to 4x faster.

Instead of predicting and typing just one word at a time like most language models, it drafts and refines entire blocks of text simultaneously.

Here’s how it works 🧵 ↓ Most large language models predict answers by guessing the single best word to say next, then the next, and so on... 🔎

It's highly capable, but not necessarily fast. The model waits to finish one word before it can think about the next.

DiffusionGemma skips the wait.

It uses "diffusion" to generate text by refining noise step by step — drafting and error-correcting whole blocks simultaneously. This makes it incredibly fast, and helpful for editing complex math and code.

New AI capabilities are coming to Google Vids, including high-quality video generation powered by Veo 3.1, available at no cost. Now, anyone with a Google account can bring stories to life from just a simple prompt or photo.

Explore more of the new features 🧵 Make custom music for your videos, powered by our Lyria 3 and Lyria 3 Pro models. Google AI Pro and Ultra subscribers can now generate everything from short 30-second clips to three-minute tracks, tailored to your video’s vibe.

We just released Gemma 4 — our most intelligent open models to date.

Built from the same world-class research as Gemini 3, Gemma 4 brings breakthrough intelligence directly to your own hardware for advanced reasoning and agentic workflows.

Released under a commercially permissive Apache 2.0 license so anyone can build powerful AI tools. 🧵↓ Gemma 4 is our most capable open model family yet:

🔵 Four versatile sizes
🔵 Up to 256K context window
🔵 Native function-calling for autonomous agents
🔵 Offline, high-quality code generation
🔵 Native multimodal support
🔵 Trained on 140+ languages
🔵 Commercially permissive Apache 2.0 license

Today @GoogleMaps is getting its biggest upgrade in over a decade. By combining our Gemini models with a deep understanding of the world, Maps now unlocks entirely new possibilities for how you navigate and explore. Here’s what you need to know 🧵 Ask Maps is a new way to get answers to your complex, real-world questions with a simple conversation. Try asking things like:

🪫 “My phone is dying — where can I charge it without having to wait in a long line for coffee?”
🎾 “Is there a public tennis court with lights on that I can play at tonight?”
🚘 "I’m headed to the Grand Canyon, Horseshoe Bend and Coral Dunes — any recommended stops along the way?"

No more sifting through research and reviews — just tap “Ask Maps” and get your questions answered with a customized map.

Ask Maps starts rolling out now in the U.S. and India on Android and iOS, with desktop coming soon.

The best of Nano Banana Pro 🍌 at the speed of Flash ⚡️

Nano Banana 2 brings studio-quality creative control and real-world reasoning to image generation.

See Nano Banana 2 in action🧵↓


1. Image generation, with advanced world knowledge.

Powered by Gemini’s real-world knowledge base and real-time web search, Nano Banana 2 can render specific subjects, like a pufferfish nest, with incredible accuracy.

Great for creating infographics, turning notes into diagrams, generating data visualizations and more.

Today, we updated Gemini 3 Deep Think to further accelerate modern science, research and engineering.

With 84.6% on ARC-AGI-2 and a new standard on Humanity’s Last Exam, see how this specialized reasoning mode is advancing research & development 🧵↓ Gemini 3 Deep Think hits benchmarks that push the frontier of intelligence.

By the numbers:
📚 48.4% on Humanity’s Last Exam (without tools)
🧩 84.6% on ARC-AGI-2 (verified by ARC Prize Foundation)
💻 3455 Elo score on Codeforces (competitive programming)

Today, we’re introducing Personal Intelligence.

With your permission, Gemini can now securely connect information from Google apps like @Gmail, @GooglePhotos, Search and @YouTube history with a single tap to make Gemini uniquely helpful & personalized to *you* ✨

This feature is launching in beta today in the @GeminiApp.

See Personal Intelligence in action 🧵 ↓ 1/ Reasoning across sources: Ever need to buy parts for your car but don't have the info handy?

Ask Gemini to “Recommend tires for my car.”

By referencing connected apps like @Gmail and Photos, it can understand your car’s make and model, and even the types of trips you take, to give recommendations of tires and info like your license plate number to make your visit to the auto shop go more smoothly.

It’s been a huge year for our image generation and editing model Nano Banana, starting with its debut in August as the top-rated image editing model in the world. Then in November, we released Nano Banana Pro. Built on Gemini 3 Pro, Nano Banana Pro uses Gemini’s state-of-the-art reasoning and real-world knowledge to visualize information better than ever before.

Amid all this momentum, it’s been incredibly fun to see the creative ways you’re actually using Nano Banana. Here are some of the most fun Nano Banana trends we’ve seen this year — which one’s your favorite? 👀🍌 1/ General photo edits

Prompt: Turn this into a night scene, lit by the moon, with dramatic shadows and moody overall lighting.

Sometimes the best ideas start with a sketch. ✍️

Here are 6 quick demos we vibe coded with simple drawings and a little help from Gemini 2.5 Pro. Try it yourself by going to gemini.google, select 2.5 Pro, tap Canvas, upload an image of your sketch, and ask, “Can you code this app?”

First up: recreate something nostalgic.

Prompt: create this app with HTML, CSS and javascript
Play interior designer before you move your furniture.

Prompt: Can you create a 3D representation of the space so I can experiment with the layout?

AI Mode is Search transformed with Gemini 2.5 at the core. It’s our most powerful AI search, with more advanced reasoning and multimodality, and the ability to go deeper through follow-up questions and helpful links to the web.

Here’s a peek at what’s coming soon to AI Mode:
🧵 📫 Personal Context

This summer, AI Mode in Labs will be able to give you personalized suggestions based on your past searches. You’ll also be able to opt in to connect other Google apps, starting with Gmail. So for example, if you usually search for and book restaurant patios, it’ll learn that you prefer outdoor seating.

It’ll be under your control, and you can choose to connect or disconnect at any time.

Play it back. Here’s a recap of (most of) our AI launches this December. Here we go ⏪ 🪩 Gemini 2.0 dropped, so it’s time to ring in the agentic era — here’s your cheat sheet for our most capable model yet.

📣 Today we’re making updates to our video and image generation models, with Veo 2 and Imagen 3, plus our newest experiment in gen AI, Whisk.

Learn more → goo.gle/3ZTcRcD Veo 2, our state-of-the-art video generation model, has:
- A better understanding of real-world physics & nuances of movement
- The ability to understand language of cinematography, like lens type & effects
- Resolution up to 4K

April 2nd is International Fact Checking Day and we’re committed to help you spot misinformation online.

Google offers a range of tools and tips to help you make sense of information online ↓ Search makes it easy for you to evaluate results.

Tapping the three dots next to a Search result can help you dig deeper and learn more about its source. bit.ly/3LtEWx8

College degrees are out of reach for many Americans, and you shouldn’t need a college diploma to have economic security. Today we’re announcing Google Career Certificates from #GrowWithGoogle to help job seekers prepare for careers in high-growth fields. goo.gle/3evINcc We will be offering 100,000 scholarships to help learners complete any of the Google Career Certificates on @Coursera so they can start new career paths in:
📈 Data Analytics
⏲️ Project Management
📲 User Experience (UX) Design
🖥️ IT Support
grow.google/certificates

And now for a Twitter thread about how to strengthen your online security. ⬇️ 1/🛡️: Set up a recovery phone number or email address and keep it updated. This keeps your account more secure by alerting you if there’s suspicious activity. g.co/security