Google Unveils Major Upgrades to NotebookLM with Gemini 3.5 and Antigravity AI
Google has announced significant enhancements to its AI tool NotebookLM, introducing advanced agentic capabilities and improved reasoning to support complex research tasks. Launched three years ago as an experimental product from Google Labs, NotebookLM has evolved into a collaborative knowledge partner for millions of users, offering tools to organize thinking, identify connections, and spark ideas. The latest upgrades, powered by Gemini 3.5 and Antigravity AI, aim to redefine how users interact with AI-driven research tools.
Enhanced Chat Experience with Gemini 3.5 and Antigravity
The revamped NotebookLM now operates on Gemini 3.5, providing more accurate and reliable information while offering greater transparency into its reasoning process. This update enables the tool to handle intricate queries with improved clarity, making it easier for users to trace how conclusions are reached. Additionally, the integration of Antigravity AI introduces new agentic capabilities, allowing the system to autonomously perform tasks such as writing and running code to support deeper analysis.
Secure Cloud Computing for Advanced Research
Each NotebookLM notebook now includes a secure cloud computer, enabling users to execute code and perform complex computations directly within the platform. This feature is complemented by over 100 curated software skills, which expand the tool’s functionality for tasks like data analysis, document summarization, and source discovery. Google emphasizes that these upgrades are designed to empower researchers, students, and professionals to tackle projects that require advanced technical expertise.
Performance Improvements and User Testing
Google conducted side-by-side evaluations of the updated NotebookLM against its predecessor, reporting an average win rate of over 65% across five core evaluation dimensions—a 15% improvement over the baseline. The system demonstrated particular strength in large document analysis, achieving a 69.9% win rate, and excelled in advanced web research and source discovery, with a 78.2% win rate. These results highlight the tool’s enhanced ability to process and synthesize information from diverse sources.
Implications for AI-Driven Research
The upgrades to NotebookLM reflect Google’s broader push to integrate AI into everyday research workflows. By combining Gemini 3.5’s language capabilities with Antigravity’s agentic features, the tool addresses gaps in traditional AI systems, such as limited reasoning depth and inability to execute code. For researchers, this means faster access to insights and the ability to perform tasks that previously required specialized software or manual intervention. However, questions remain about how these improvements will scale across industries and whether they will set a new standard for AI research tools.
What’s Next for NotebookLM?
As Google continues to refine its AI ecosystem, the success of NotebookLM’s upgrades could influence future developments in agentic AI. The company has not yet outlined specific plans for expanding the tool’s capabilities, but the focus on secure cloud computing and advanced reasoning suggests a long-term commitment to AI-driven research. Users will likely be watching closely to see how these features evolve and whether they can address broader challenges in information retrieval and analysis.