Hey everyone!

Google I/O 2025 wrapped up on Wednesday and left us with a massive amount of AI news. They announced a number of new models as well as a flurry of new AI features. From supercharging Search and Google Workspace to unveiling a new AI video tool, there has been so much information to sift through. We’ll be exploring all the major AI updates that were announced at Google I/O 2025. Let’s jump in!

Gemini 2.5 Family of Models Arrives

Headlining Google I/O was the official unveiling of Gemini 2.5, the latest version of Google's AI model family. This update brings a significant leap forward, aiming for increased speed, intelligence, and performance across the board. 

• Gemini 2.5 Pro: Once again Gemini 2.5 Pro is the star of the show. Google announced a series of updates including enhanced reasoning capabilities and a new “Deep Think” mode. Deep Think was designed to allow the model to consider multiple answers to your query at once and will significantly improve code generation and math problems. While 2.5 Pro is already at the top of the LMSYS leaderboards, the incorporation of LearnLM has made 2.5 Pro the best model for learning.

• Gemini 2.5 Flash: Google also released the latest Flash model as a lightweight alternative to 2.5 Pro. It’s designed to be faster and cheaper than other models and has consistently hit that goal. Flash is already performing extremely well on the leaderboards, placed just below the latest 2.5 Pro model. 

New AI Tools in Google Search 

Google Search needs no introduction, it's been a cornerstone of the internet for decades and it received several new AI features at Google I/O. Unsurprisingly, Google has further embedded their Gemini models into Search with a focus on new modes and multimodal capabilities. 

• AI Mode: This new end-to-end AI Search experience was built off of my favorite feature from last year’s Google I/O, the AI Overview. AI Mode is meant to find and collect relevant information from across the web to provide direct answers to your questions. The goal for this tool is that it will replace (or limit) the need to piece together an answer from multiple sources. This new AI Mode is powered by Gemini 2.5 Pro (which is still topping the LMSYS Leaderboards) and will allow you to give complex, multi-stage prompts that require reasoning. The new mode works by breaking your query into subtopics and simultaneously searching each element of your larger question. The goal of this new feature is to help get reliable answers faster and more efficiently. Google added that this is just the beginning for AI Mode and that new features will be released regularly.

• Live Search: This new Search feature will expand on Google Lens by allowing you to capture your surroundings with your camera and directly ask Google questions about what you're seeing. Now, you can simply point your camera at something and ask Google about it. Whether you need to identify an object, get instructions, understand a diagram, or even find information about a place you’re looking at, Live Search will help make this process quick and easy. 

New AI Image and Video Models

For anyone interested in AI-powered content creation, you’ll be thrilled to know that Google has unveiled three major updates; Veo 3, Imagen 4, and Flow. 

• Veo 3: Video generation with audio included? Yes please! Veo 3 is Google’s latest AI video generation model and it comes with some major leaps forward. It now included 4k quality output with far greater realism than any previous Veo models. The most significant upgrade has to be the inclusion of generated audio. Based on the videos Google has shared, the audio quality is remarkably high and includes dialogue, music, and sound effects. It helps turn ordinary AI generated videos into an immersive experience. 

• Imagen 4: Google’s flagship image generation model has received a 4th version that pushes the limits of AI generated content. This latest model is 10x faster than Imagen 3 and comes with a number of improvements. These include better text creation, 2k resolution, and significantly better detail for textures. You can already start creating with Imagen 4 using the Gemini app. 

• Flow: This was the most surprising media generation release that we got from Google I/O. Flow is a new “AI filmmaking tool that lets you seamlessly create cinematic clips, scenes and stories” by utilizing Imagen 4, Veo 3, and Gemini. This tool was designed to allow you to describe what you want to create using natural language. Ultimately turning you into a filmmaker in just minutes. Flow is currently available to all Gemini Pro and Ultra users.

New AI Tools and Updates

Aside from Search and Media creation-related updates, Google unveiled a number of new and updated AI tools. Let’s dive in!

• Stitch: Google announced a new prototype AI tool that is meant to take vibe coding to the next level. Stitch is a tool that allows you to turn simple prompts and images into high-level UI designs and frontend code. Whether you’re creating a small personal website or need to redesign the homepage of your business, Stitch will allow you to customize your UI until it's perfect. Stitch designs are meant to be exported and can be seamlessly integrated into Figma or your codebase. 

• NotebookLM: Google’s research tool has received several big updates at Google I/O. First and foremost is an update to Audio Overviews. These great summaries have been updated to allow for greater customization and personalization. Now you can share custom instructions about the overview and even control its length. They are also rolling out new Video Overviews that will build off Audio Overviews by including generated graphics that follow the conversation. You can learn more about NotebookLM through our in-depth tutorial. 

• Gemini Ultra Plan: Google has announced a new Gemini subscription plan called “Ultra” that brings all the latest and greatest Gemini products into one place. This new plan will cost a surprising $249.99/month but is meant to bring all of Google’s premiere features together. This plan includes the following:

  • The highest usage limits for Gemini.

  • Google’s new AI filmmaking tool Flow. 

  • Better models and higher usage limits in NotebookLM.

  • Google’s new Project Mariner, Google’s new prototype agentic assistant.

  • Youtube Premium with all its usual benefits.

  • 30 TB of storage within Google Drive, Photos, and Gmail. 

  • New Gemini Chrome features that allow Gemini to interact with your Chrome and help complete tasks.

• SynthID Checker: This new platform will help you identify AI generated content for Google specifically. The hope is that with AI generated content being increasingly more common and hard to identify, Google’s new portal will quickly allow anyone to know whether something was created by their models. While this is a good starting place, I do think AI providers may need to go a step further to ensure that AI generated content can be identified when needed. 

• Google Workspace: Google has announced a series of updates coming to Google Workspace. These include Imagen 4 capabilities across Slides, Docs, and more. Search Grounding has also been added to Docs to help make writing more efficient. Some of the biggest changes have come to Gmail where now you can receive email summaries for long email threads, inbox cleanup which allows you to filter large amounts of emails with Gemini, and personalized “smart” replies that help speed up response times. 

Conclusion:

The overriding theme from Google I/O 2025 is undeniable: AI is no longer just a product category; it's becoming the core of Google's entire ecosystem. Whether you’re using their next generation search features or creating short films with Flow, AI has become an integral element of Google’s products. Personally, I’m excited by this. Everything they’ve shared speaks of a company that is making AI more accessible without hiding it behind the scenes. Don’t want to use AI Overviews or AI Mode when searching on Google? Well you don’t have to. But everyone who does can now easily access it. 

While this newsletter covered a ton of new developments, I’ll be exploring these updates in greater detail in the coming weeks. If you want more info today, check out this NotebookLM project that breaks down everything that was covered during Google I/O 2025.