Google I/O: AI Subscription & Agent Gemini
- – Google's annual I/O developer conference kicked off with CEO Sundar Pichai outlining a vision for artificial intelligence that is more intelligent, agent-like, and personalized.The overarching goal, according...
- Hassabis stated that Gemini 2.5 Flash and Pro are the cornerstones of this new AI paradigm.
- A new "Reasoning" model,Deep Think,is incorporated into Gemini 2.5.
Google I/O 2025: Gemini AI Takes Center Stage with New Features, Subscription Model
Table of Contents
- Google I/O 2025: Gemini AI Takes Center Stage with New Features, Subscription Model
- Google I/O 2025: gemini AI Takes Center Stage
- What was the main focus of Google I/O 2025?
- What is Gemini 2.5?
- What are the key features of Gemini 2.5?
- How is Gemini 2.5 being integrated into Google’s products?
- what is the AI Ultra subscription, and what does it offer?
- How does the new AI-powered Search work?
- What advancements were announced in image and video generation?
- What are the updates coming to other Google services?
- Can you provide a summary of the new technologies announced?
MOUNTAIN VIEW, Calif. – Google’s annual I/O developer conference kicked off with CEO Sundar Pichai outlining a vision for artificial intelligence that is more intelligent, agent-like, and personalized.The overarching goal, according to Pichai, is for AI to provide better, more accurate, and detailed answers, while also proactively performing tasks in the user’s best interest. Demis Hassabis, CEO of Google DeepMind, envisions a future where AI models possess comprehensive world knowledge, mirroring the capabilities of the human brain.
Gemini 2.5: Foundation for the Future of AI
Hassabis stated that Gemini 2.5 Flash and Pro are the cornerstones of this new AI paradigm. Google aims to enhance these models and broaden their accessibility. Gemini 2.5 Flash is now integrated into the Gemini app and available to developers through Google AI Studio and Vertex AI. Gemini 2.5 Pro is expected to follow soon, building upon a preliminary version already released to developers. Google reports that Gemini 2.5 Pro leads the WebDev Arena with an ELO of 1420, excelling across all LMarena categories where AI models are competitively evaluated.
Gemini’s Enhanced Reasoning and Browser Control
A new “Reasoning” model,Deep Think,is incorporated into Gemini 2.5. This model is designed to logically connect data rather than simply reproducing content. Both Gemini models will soon feature native audio generation, capable of replicating a speaker’s key, accent, and style. The models are also being developed as agents capable of controlling web browsers, a project previously known as Project Mariner. This functionality is being integrated into the Gemini API and vertex AI. Google is also adopting Anthropic’s Model Context Protocol (MCP), enabling Gemini to gather information and create lists by making websites more “readable” for AI agents.
AI Ultra subscription: A $250 “VIP Pass”
Project mariner is currently available to subscribers of Google’s new AI Ultra subscription in the U.S.This premium service, dubbed a “VIP pass,” costs $250 per month. It includes expanded usage limits for all AI services, granting access to Deep Think, Flow, and whisk, which transforms images into short animations. The subscription also provides 30TB of storage for photos and documents.The existing AI Premium plan is being renamed AI Pro, offering continued and expanded access to features like Flow.
AI-Powered Search and Shopping
Google is expanding its AI Mode in search within the U.S.,building upon its initial testing phase in Google Labs.This AI Mode leverages reasoning capabilities to provide more comprehensive answers to complex queries. google’s Knowledge Graph, a vast database of collected information, powers this feature. AI is also being integrated into Google Shopping,enabling features such as virtual clothing try-ons and agent-assisted checkout,features previously tested by Google.
Image and Video Generation Advancements
Images 4 and VEO 3 represent google’s latest advancements in video and image generation.VEO 3 will soon be capable of generating native audio, including text and ambient noise. Flow aims to facilitate the creation of longer films with enhanced control over characters and styles. Google showcased these capabilities with a video at the I/O keynote, depicting a wild-west city populated by parrots and a T-Rex constructed from clamping modules.
Updates are also coming to Canvas, Google’s AI-powered sketching tool, Deep Research (a chatbot for scientific tasks), Gmail, and Google Meet. Gmail will gain the ability to generate email responses,while Google Meet will offer synchronous translations in select languages.
Google Beam, Android XR, and Gemini Live
Google Beam, a new AI-driven 3D interaction platform, was unveiled. It offers simultaneous translation of conversations in the speaker’s own voice. Android XR aims to bring Gemini to smart glasses and headsets. gemini Live, available on Android and iOS, allows users to interact with the AI in real-time using their device’s camera.A demonstration video showed gemini assisting with a bicycle repair by calling a shop for a spare part.
Google I/O 2025: gemini AI Takes Center Stage
What was the main focus of Google I/O 2025?
The primary focus of Google I/O 2025 was artificial intelligence, with a strong emphasis on the Gemini AI model. google CEO Sundar Pichai outlined a vision for AI that is more intelligent, agent-like, and personalized. The overarching goal is too provide better, more accurate, and detailed answers, while also proactively performing tasks in the user’s best interest.
What is Gemini 2.5?
Gemini 2.5 is the next iteration of Google’s AI model, described as the foundation for the future of AI. It includes enhancements to its capabilities with the integration of models like Gemini 2.5 Flash and Pro.Gemini 2.5 Pro is reported to lead the WebDev Arena with an ELO of 1420.
What are the key features of Gemini 2.5?
Gemini 2.5 boasts several key features:
Enhanced Reasoning: Incorporates a new “reasoning” model, Deep Think, designed to logically connect data, rather than simply reproducing content.
Native Audio Generation: Both Gemini models will soon feature native audio generation,capable of replicating a speaker’s key,accent,and style.
Browser Control: Gemini models are being developed as agents capable of controlling web browsers, a project previously known as Project Mariner.
Model Context Protocol (MCP) adoption: Google is adopting Anthropic’s Model Context Protocol (MCP), enabling Gemini to gather facts and create lists by making websites more “readable” for AI agents.
How is Gemini 2.5 being integrated into Google’s products?
Gemini 2.5 Flash is now integrated into the Gemini app and available to developers through Google AI Studio and Vertex AI. The browser control functionality (Project Mariner) is being integrated into the Gemini API and Vertex AI. Gemini is also being incorporated into various other Google services like Search and shopping.
what is the AI Ultra subscription, and what does it offer?
The AI Ultra subscription is a premium service, also known as a “VIP pass,” priced at $250 per month.It offers:
Expanded usage limits for all AI services.
Access to Deep Think, Flow, and whisk (transforms images into short animations).
30TB of storage for photos and documents.
How does the new AI-powered Search work?
Google is expanding its AI Mode in Search within the U.S., building upon its initial testing phase in Google Labs. This AI Mode leverages gemini’s reasoning capabilities to provide more complete answers to complex queries. Google’s Knowledge Graph powers this feature.
What advancements were announced in image and video generation?
Google announced advancements in image and video generation with Images 4 and VEO 3.Notably:
VEO 3: Will soon generate native audio, including text and ambient noise.
Flow: Aims to facilitate the creation of longer films with enhanced control over characters and styles.
What are the updates coming to other Google services?
notable updates are coming to several other Google services:
Canvas: Google’s AI-powered sketching tool will receive updates.
Deep Research: A chatbot for scientific tasks will be improved.
Gmail: Will gain the ability to generate email responses.
* Google Meet: Will offer synchronous translations in select languages.
Can you provide a summary of the new technologies announced?
Here’s a concise summary of the new technologies and their functionalities:
| Feature | Description | Integration |
| ———————- | ————————————————————————————————– | ——————————————— |
| Gemini 2.5 (Flash & Pro) | Next-generation AI models with enhanced reasoning and audio generation. | Gemini app, Google AI Studio, Vertex AI |
| Deep Think | New reasoning model for logically connecting data. | Gemini 2.5 |
| Native Audio Generation| Models capable of replicating speaker’s voice, accent, and style. | Gemini 2.5 |
| Browser Control | AI agents capable of controlling web browsers (Project Mariner). | Gemini API, Vertex AI |
| Google Beam | New AI-driven 3D interaction platform with simultaneous translation. | Self-reliant Platform |
| Android XR | Bringing Gemini to smart glasses and headsets. | Android Devices |
| Gemini Live | Real-time interaction with AI using device’s camera. | Android & iOS |
| Images 4 & VEO 3 | Advancements in video and image generation including the ability to create native audio on VEO 3.| image/Video Generation |
| AI Ultra Subscription | Premium AI service with expanded usage limits and access to advanced features. | U.S. Users |