Google has announced updates to its Gemini AI platform at Google I/O, introducing features that could transform how search and marketing professionals analyze data and interact with digital tools.
The new capabilities focus on enhanced reasoning, improved interface interactions, and more efficient workflows.
Gemini 2.5 Models Get Performance Upgrades
Google highlights that Gemini 2.5 Pro leads the WebDev Arena leaderboard with an ELO score of 1420. It ranks first in all categories on the LMArena leaderboard, which measures human preferences for AI models.
The model features a one-million-token context window for processing large content inputs, effectively supporting both long text analysis and video understanding.
Meanwhile, Gemini 2.5 Flash has been updated to enhance performance in reasoning, multimodality, code, and long context processing.
Google reports it now utilizes 20-30% fewer tokens than previous versions. The updated Flash model is currently available in the Gemini app and will be generally available for production in Google AI Studio and Vertex AI in early June.
Gemini Live: New Camera and Screen Sharing Capabilities
The expanded Gemini Live feature is a significant addition to the Gemini ecosystem, now available on Android and iOS devices.
Google reports that Gemini Live conversations are, on average, five times longer than text-based interactions.
The updated version includes:
- Camera and screen sharing capabilities, allowing users to point their phones at objects for real-time visual help.
- Integration with Google Maps, Calendar, Tasks, and Keep (coming in the next few weeks).
- The ability to create calendar events directly from conversations.
These features enable marketers to demonstrate products, troubleshoot issues, and plan campaigns through natural conversations with AI assistance.
Deep Think: Enhanced Reasoning for Complex Problems
The experimental “Deep Think” mode for Gemini 2.5 Pro uses research techniques that enable the model to consider multiple solutions before responding.
Google is making Deep Think available to trusted testers through the Gemini API to gather feedback prior to a wider release.
New Developer Tools for Marketing Applications
Several enhancements to the developer experience include:
- Thought Summaries: Both 2.5 Pro and Flash will now provide structured summaries of their reasoning process in the Gemini API and Vertex AI.
- Thinking Budgets: This feature is expanding to 2.5 Pro, enabling developers to manage token usage for thinking prior to responses, which impacts costs and performance.
- MCP Support: The introduction of native support for the Model Context Protocol in the Gemini API allows for integration with open-source tools.
Here are examples of what thought summaries and thinking budgets look like in the Gemini interface:
Gemini in Chrome & New Subscription Plans
Gemini is being integrated into Chrome, rolling out to Google AI subscribers in the U.S. This feature allows users to ask questions about content while browsing websites.
You can see an example of this capability in the image below:
Google also announced two subscription plans: Google AI Pro and Google AI Ultra.
The Ultra plan costs $249.99/month (with 50% off the first three months for new users) and provides access to Google’s advanced models with higher usage limits and early access to experimental AI features.
Looking Ahead
These updates to Gemini signify notable advancements in AI that marketers can integrate into their analytical workflows.
As these features roll out in the coming months, SEO and marketing teams can assess how these tools fit with their current strategies and technical requirements.
The incorporation of AI into Chrome and the upgraded conversational abilities indicate ongoing evolution in how consumers engage with digital content, a trend that search and marketing professionals must monitor closely.
