Google wants Gemini to work while you sleep: Check biggest new features announced at I/O 2026 | Mint

Google took the wraps off a few of its biggest AI updates at I/O 2026, together with the Gemini 3.5 Flash mannequin, Gemini Omni, Gemini Spark, and the new Antigravity platform. However maybe the biggest change the corporate showcased this 12 months was remodeling Gemini from a chatbot expertise into an agentic one, not simply with Gemini Spark but additionally by including devoted AI brokers contained in the app.

Here is all the things Google announced for Gemini this 12 months.

1) Gemini Spark:

Google announced a new 24/7 private AI agent named Gemini Spark, designed to work within the background and actively full duties on behalf of customers.

The new platform is powered by the Gemini 3.5 Flash mannequin, which Google additionally unveiled at I/O 2026, while benefiting from Antigravity 2.0. Spark can proceed working even when customers lock their telephones or shut their laptops. The agentic AI instrument can also be deeply built-in with Google Workspace apps like Gmail, Docs, and Slides, permitting it to take extra complicated actions throughout a consumer’s digital life.

With Gemini Spark, customers can:

Set recurring triggers: Instruct the agent to mechanically parse month-to-month bank card statements and flag hidden subscription charges.

Educate customized abilities: Have Spark monitor inboxes for emails from a toddler’s college, extract important deadlines, and compile a each day digest.

Execute full workflows: Customers can ask the agent to synthesise uncooked assembly notes from emails and chats, draft a refined Google Doc, and write a companion e mail to kick off a mission.

Google says Spark requires specific permission from customers earlier than making high-stakes choices like spending cash or sending emails.

2) Each day Temporary:

Google can also be introducing a new AI agent contained in the Gemini app known as “Each day Temporary”. The agent is designed to give customers a “personalised morning digest that is designed to be your first cease every single day.”

The new agent is an opt-in expertise, that means customers could have to explicitly allow it earlier than utilizing it. The characteristic connects with numerous apps within the background and compiles info right into a skimmable, prioritised briefing based mostly on customers’ objectives while additionally suggesting fast subsequent steps.

3) Gemini app redesigned:

Google has additionally redesigned the Gemini interface throughout each the app and net variations of the platform. The new design, known as “Neural Expressive”, brings fluid animations, vibrant colors, refreshed typography, and haptic suggestions.

Google has additionally baked Gemini Reside extra immediately into the Gemini app to make switching from typing to voice interactions simpler. In the meantime, the corporate showcased regional dialect help, together with some Indian languages, which can quickly be obtainable inside Gemini Reside.

4) Gemini Omni:

Google can also be including its newest “world mannequin”, Gemini Omni, to the Gemini app for AI video technology and enhancing. The corporate says Omni can take combos of textual content, pictures, and movies as inputs and rework them into cinematic-quality movies.

Customers can add footage immediately from their digicam roll and use easy textual content prompts to apply cinematic zooms, swap backgrounds, or use built-in templates with a single faucet. They will even create customized AI avatars of themselves to place inside generated movies.

The primary model on this sequence, Gemini Omni Flash, is rolling out now to Google AI Plus, Professional, and Extremely subscribers.

5) Gemini app for macOS:

Google not too long ago rolled out its Gemini app for macOS and at I/O 2026, the corporate announced that the app will get Gemini Spark help later this summer time. This might enable the AI assistant to handle native recordsdata and automate workflows natively on customers’ machines.