The Google I/O 2026 keynote is now in the books, and Google had a lot to share. The entire keynote event lasted nearly two hours and was full of new product and feature announcements.

With all that was unveiled at I/O, AI was clearly Google's focus, with news about Gemini taking center stage.

Here's everything Google announced about Gemini at I/O 2026.

Gemini 3.5

Google CEO Sundar Pichai introduces Gemini 3.5 Flash at Google I/O 2026. Credit: Google / YouTube

Google announced a new family of AI models, Gemini 3.5, at I/O 2026. The first model to launch is Gemini 3.5 Flash, with Gemini 3.5 Pro rolling out next month.

According to Google, Gemini 3.5 Flash is four times faster than other frontier models in output tokens per second (TPS), which measures a model's response time. Google says it also outperforms most frontier models on key benchmarks, including Gemini 3.1 Pro.

During the keynote, Google announced that Gemini 3.5 Flash is now the default model powering the Gemini app and Google search's AI Mode.

Gemini Omni

Omni is the new world model from Google DeepMind. Credit: Google

Gemini 3.5 wasn't the only new family of models announced at I/O 2026. Google also announced Gemini Omni, a new world model. At the end of the keynote, Google DeepMind CEO Demis Hassabis described Omni as a pivotal step toward artificial general intelligence, or AGI.

The first model in the family, Gemini Omni Flash, was showcased at the event.

According to Google, unlike text-to-video AI models, Gemini Omni is multi-modal in both input and output. Users can input text, audio, images, and video into the model when crafting their prompt and, utilizing Gemini's "real-world knowledge," Omni will generate realistic and scientifically accurate content.

Gemini Omni Flash rolls out today to paid Google AI Plus, Pro, and Ultra subscribers within the Gemini app and Google Flow. Gemini Omni Flash will also launch in YouTube Shorts and the YouTube Create app at no cost to users later this week.

Gemini Spark

Gemini Spark is easily the most ambitious product revealed at I/O. Google is calling Spark "your personal AI agent" and integrates with Google's suite of products, and eventually with more than 30 third-party tools via MCP, such as Adobe, Dropbox, and Uber.

Acting as an AI agent, Spark can pull together relevant emails from your inbox and files from Docs in order to craft an update for your boss. 

Notably, Spark runs entirely within the cloud and doesn't require any hardware.

Google AI Ultra subscribers in the U.S. will see Gemini Spark roll out inside Gmail and Chat in the next week. 

Google AI subscription changes

Google's new AI subscription tiers. Credit: Google

Speaking of the Google AI Ultra subscription, the company announced that it's lowering the starting price for its Ultra subscription to $200 per month and introducing a lower-cost $99 per month Ultra tier.

Previously, users had to pay a whopping $250 per month to access the latest Gemini features such as AI Inbox.

Neural Expressive

Credit: Google

Finally, the Gemini app is getting a new design language, which Google calls Neural Expressive. The new design language includes more fluid animations, haptic feedback, and vibrant colors, per Google. Neural Expressive also integrates the Gemini Live voice experience into the core UI, meaning users no longer have to switch back and forth.

Google says it's rolling out globally on Gemini's desktop, Android, and iOS apps.

Are you an Apple superfan? Enter Mashable’s Big Guessing Game to win prizes.