Google has announced a major set of upgrades for its Gemini AI app at the I/O 2025 event, and these changes clearly show how serious the company is about competing with OpenAI’s ChatGPT and Apple’s Siri. With features like real-time video understanding, deep research tools, and tighter integration with Google’s popular apps, Gemini is evolving into much more than just a chatbot.
Starting this week, both Android and iOS users will be able to use Gemini Live with full camera and screen-sharing support. This means you can now talk to Gemini using your voice while showing it what you see through your phone’s camera — all in near real-time. Powered by Google’s Project Astra, the feature allows for natural, flowing conversations where the AI can see and understand your environment. For example, if you are exploring a new city, you can point your camera at a building and ask about its architecture or history, and Gemini will instantly respond with relevant information.
This feels like a big step forward. It is not just about chatting anymore — it is about making AI observe and react in real time, just like a smart assistant should. Google says the tool will soon go even further by connecting with other services like Google Maps, Google Calendar, and Google Tasks. So imagine pointing your phone at a flyer and saying, “Add this event to my calendar” — and it just does it.
Competing With the Best
It is no secret that Gemini is Google’s answer to ChatGPT. With AI becoming the core of how users interact with their devices, Google cannot afford to fall behind. According to the company, Gemini now has over 400 million monthly active users, and these new updates are clearly designed to push that number even higher.
To match OpenAI’s subscription model, Google is now offering two paid tiers:
- Google AI Pro, which is a rebrand of its $20/month Gemini Advanced plan.
- Google AI Ultra, a new $250/month plan aimed at power users and professionals.
The Ultra plan gives users access to Google’s newest models, higher usage limits, and early access to experimental features. This pricing shows that Google is targeting business users, researchers, and creators who want advanced AI tools for work or content creation.
Users on these paid plans (in the U.S., with English set as their language) will also see Gemini integrated directly into Chrome. This means you can highlight content on a web page and ask Gemini to summarize or explain it, right inside your browser. That alone could change how we use the internet.
Smarter Research and Better Media Tools
One of the most impressive updates is to Deep Research, Gemini’s research-focused AI assistant. You can now upload your own PDFs and images, and Gemini will combine that with public data to generate detailed, personalized reports. Google also plans to let users connect Google Drive and Gmail directly to Deep Research, so the AI can pull insights from your own files and emails.
On the creative side, Google is rolling out its latest image model Imagen 4 for free users, promising better-quality visuals and text output. Meanwhile, AI Ultra subscribers will get access to Veo 3, a powerful new video model that not only creates visuals but also generates sound that matches the scenes — which is a big step for creators looking for more realistic AI-generated content.
Lower Latency, Smarter Learning
Google is also switching the default AI model in Gemini to Gemini 2.5 Flash, which they say gives faster and more accurate results. That means shorter waits and more helpful answers, something everyone will appreciate.
And for students, there is a new quiz system that creates personalized questions based on your weak areas. If you answer something wrong, Gemini will create new quizzes and study plans to help you improve. This could be a valuable tool for students trying to prepare for exams or learn difficult topics more efficiently.
Final Thoughts
With these updates, Google is turning Gemini into an all-in-one AI companion, not just for chatting, but for productivity, learning, content creation, and real-world assistance. If Google continues at this pace, Gemini could become one of the most useful AI tools available.
That said, the success of features like Gemini Live and Deep Research will depend on how well they actually perform outside of demos. Real-world use will reveal how natural the AI feels and whether it truly understands complex, visual inputs the way Google claims.
Still, it is clear Google is betting big on Gemini, and these latest upgrades show just how far they are willing to go to stay ahead in the AI race.
