Google Gemini already offers a host of useful capabilities. From generating text and creating images to live conversations, deep research, and analyzing files, Google’s AI has proven itself a strong contender in the AI field. At Google I/O 2025 on Tuesday, the company revealed a slew of new and improved features now available with its AI assistant.
Also: How to watch Google I/O 2025 (and why you shouldn’t miss it)
1. New Google AI Pro and Ultra plans
First up are two new Google AI subscriptions that offer more features but naturally come with their own price tags.
The first plan is known as Google AI Pro, which is actually the same AI Premium plan that’s been around for a while just with a new name. Still priced at $20 per month, AI Pro offers the same AI features available with the free version of Gemini but adds higher rate limits and special features. AI Pro also includes the Gemini app formerly known as Gemini Advanced, along with products like NotebookLM and the new Flow AI video editor. Those two features will reach AI Pro subscribers in the US first and then expand to other countries.
College students in the US, the UK, Brazil, Indonesia, and Japan can get a free school year of Google AI Pro.
If you need more power and features and are willing to spend the big bucks, there’s also a Google Al Ultra plan. This one offers the most powerful models, the highest rate limits, and early access to experimental Al features.
As one example, the Ultra plan will grant you early access to Agent Mode, a new desktop-based agentic tool that will carry out tasks for you. Just describe your request or question; in response, the agent browses the web, conducts its own research, and integrates with your Google apps to tackle complex, multi-step tasks from start to finish.
The Ultra plan costs a hefty $250 a month, though first-time subscribers get 50% off for the first three months.
2. Gemini Live
Next is Gemini Live, the handy chat mode in which you carry on a back-and-forth voice conversation with the AI. Previously, only Android users could share their screen or camera view and ask Gemini questions about it. Now, Google is expanding this feature so that Android and iOS users alike will be able to use the camera and screen sharing.
Also: Google’s viral NotebookLM AI tool gets its own Android app – what you can do with it
To try this, open the Gemini app on your iPhone or Android device and tap the Gemini Live icon to the right of the prompt. The camera icon at the bottom lets you aim your phone at any object or scene and ask Gemini to describe it or answer questions about it. The second icon allows you to share any screen on your device for Gemini to analyze.
There’s more: In the coming weeks, Gemini Live will work with other Google apps and services, including Google Maps, Calendar, Tasks, and Keep. This means you’ll be able to ask Gemini Live to perform such tasks as creating a calendar appointment or providing directions to your next destination.
3. Imagen 4 image generation
Previously, Google used its Imagen 3 model to generate images based on your descriptions. Now, the company has upgraded to Imagen 4, which it claims will offer faster performance, more lifelike details, and better text output. Anyone will now be able to try Imagen 4 via the Gemini mobile app.
4. Veo 3 video generation
Also getting an upgrade is Gemini’s Veo video generator. Moving up from Veo version 2, Veo 3 offers native audio generation with support for dialogue between characters, background noises, and sound effects. As Google describes it, you can now add anything from bustling city sounds to the rustle of leaves to character dialogue just from your text descriptions. The main barrier here is that Veo 3 will be available only to Google AI Ultra subscribers in the US.
5. Canvas enhancements
Google’s Canvas tool offers you an interactive and collaborative workspace in which you can create code, design web pages, and devise other visual content, with the results appearing side-by-side in real-time. Using the latest Gemini 2.5 model, Canvas promises to be more intuitive and powerful, according to Google.
Also: Google Beam is poised to bring 3D video conferencing mainstream
You can create interactive infographics, quizzes, and podcast-style Audio Overviews in any one of 45 languages. With Gemini 2.5 Pro’s coding skills, Canvas is now more adept at converting your ideas into actual code, thereby helping you develop full applications.
6. Interactive quizzes
Trying to learn a complicated new subject? Gemini may be able to help. You can now ask the AI to create a quiz on your topic of interest. In response, Gemini challenges you with a series of questions designed to expand your knowledge. As you answer each question, the AI will tell you how you’re doing and focus on any areas that need special attention. This feature is now rolling out to all Gemini users on desktop and mobile devices.
7. Gemini-in-Chrome
As of Wednesday, Gemini will begin popping up in Chrome on the desktop in both Windows and MacOS. Here, you’ll be able to ask Gemini to analyze or answer questions about your current web page. Down the road, the AI will also work across multiple tabs and even launch different websites for you.
Sounds helpful, but access will be limited. Gemini-in-Chrome will be available only to Google Al Pro and Google Al Ultra subscribers in the US who use English as their language in the browser.
Also: Meet Gemini-in-Chrome, your new AI browsing assistant – here’s who gets to use it
8. Deep Research
Finally, Gemini’s Deep Research mode is an agentic tool that can conduct online research for you and present the results in a detailed report, all on its own. Previously, Deep Research was only able to consult websites for the information you needed. Now, it can also check out your own PDFs and images. This means you could tell Gemini to include trends and topics that have already been captured in your own personal or work files.
In one example cited by Google, a market researcher could upload internal sales figures stored in a PDF to cross-reference with public market trends. In another example, an academic researcher could tell Gemini to consult downloaded journal articles to add to a review of online literature. As one more item, Google said that it plans to integrate Deep Research with Google Drive and Gmail to expand the number of sources available.
Also: Google gives AI-generated video a voice with Veo 3 – how to try it
Whew, that’s a lot to unpack. But with AI increasingly impacting both individuals and organizations, Google is showing that it’s trying to stay competitive. And even with the pricey, new Ultra subscription, there’s enough here for free Gemini users and AI Pro subscribers to try and see if and how they can take advantage of the latest developments.
Want more stories about AI? Sign up for Innovation, our weekly newsletter.