Google’s NotebookLM has received its most significant update to date, transitioning to the Gemini 3.5 model and integrating “Antigravity” support to automate complex research workflows. According to Google, this upgrade enhances the platform’s core capabilities across five dimensions, including accuracy, document analysis, and multilingual support, while allowing users to execute code directly within the interface.
How the Gemini 3.5 Upgrade Changes Research
Google has shifted NotebookLM from the older Gemini 3.1 branch to the newer Gemini 3.5 Flash model. This transition, which debuted at Google I/O earlier this year, is designed to provide faster and more efficient processing for heavy data tasks.
The performance gap between the two models is quantifiable. Based on Google’s internal side-by-side evaluations, the updated NotebookLM achieved a 65 percent win rate over the previous iteration. The company measured this success across five specific metrics:
- Accuracy and Quality
- Multilingual Support
- Large Document Analysis
- Document Creation
- Advanced Research
By moving to the Flash model, Google aims to provide users with higher quality outputs while optimizing the underlying computational costs.
What is the New “Cloud Computer” Integration?
One of the most notable additions is the integration of what Google calls a “cloud computer,” which leverages Antigravity to run code. This feature allows NotebookLM to move beyond simple text summarization and actively perform research-related tasks.
Users now have access to over 100 software skills built directly into the platform. These skills are designed to replace the need to switch between multiple applications, enabling users to build automated workflows within their notebooks. By allowing the AI to write and execute code, Google is positioning NotebookLM as a functional workspace rather than just a passive reading tool.
Pro Tip: If you frequently jump between different apps to organize your research, the new software skills in NotebookLM are designed to consolidate those tasks into a single, cohesive workflow.
Future Trends in AI Research Tools
Since its launch in 2023, NotebookLM has remained an outlier in Google’s product lineup by avoiding the "shut down" fate of many experimental tools. The recent push toward agentic capabilities—where the AI doesn’t just answer questions but performs actions—suggests a shift in how research software will function moving forward.
As models like Gemini 3.5 Flash become standard, the barrier to entry for complex data analysis continues to drop. We are moving toward an era where “doing research” involves delegating the heavy lifting of code execution and document synthesis to an AI agent that operates within a unified digital environment.
Frequently Asked Questions
Is NotebookLM still free to use?
Google’s recent updates focus on performance and capability, and the platform remains a key part of the company’s generative AI offerings.
What is the main advantage of the Gemini 3.5 Flash upgrade?
According to Google, the primary benefits are faster processing speeds and improved output quality compared to the 3.1 branch, with a 65 percent win rate across core evaluation dimensions.
Can I run my own code in NotebookLM?
The new update introduces a "cloud computer" feature that uses Antigravity to write and run code for you, specifically to support your research goals.
What kind of documents can I analyze?
NotebookLM supports a wide range of file types and has streamlined web source integration, allowing you to analyze specific webpages alongside your uploaded documents.
Are you using NotebookLM to streamline your research projects? Let us know your favorite new workflow in the comments below, or subscribe to our newsletter for more updates on the latest in AI productivity tools.
