Google NotebookLM Updates with Gemini 3.5 Flash and Antigravity: More Intelligence, Less Friction

Google has updated NotebookLM, its generative AI platform for document analysis, with a substantial update that integrates the new Gemini 3.5 Flash model and the Antigravity feature. This initiative confirms Google's strategy to refine its AI tools for the experienced user, aiming for a tangible optimization of research and creation workflows.

The integration of Gemini 3.5 Flash, previously introduced at Google I/O, represents the core of the upgrade. This model promises significantly faster and more efficient processing, allowing NotebookLM users to analyze specific sources - documents and web pages - with greater speed and optimized resource consumption. Google highlighted that the transition to Flash enables significant token cost savings while maintaining or even improving output quality compared to previous models.

Internal tests conducted by Google on "five key evaluation dimensions" (Accuracy and Quality, Multilingual Support, Bulk Document Analysis, Document Generation, and Advanced Search) showed a 65% superiority of the new model.

The most intriguing addition, however, is the introduction of Antigravity. This feature equips NotebookLM with a true internal 'cloud computer', enabling it to write and execute code autonomously to support the user's research goals. Antigravity includes a suite of over 100 'software skills', providing the capability to build complex and automated workflows directly within NotebookLM. This eliminates the need to constantly switch between different applications. This capability transforms NotebookLM from a simple analysis tool into a complete AI agent, capable of proactively processing and synthesizing information.

The update also extends support to new file types and enhances the integration of web sources, making NotebookLM even more versatile in data ingestion. Since its debut in 2023, NotebookLM has positioned itself as one of Google's leading tools in the generative AI landscape, and this massive injection of technology strengthens its vocation as an environment for those requiring in-depth and structured analysis, without succumbing to the purely ambitious specifics that often afflict the sector.

Users can explore the details of the update and its implications for AI workflows by consulting the official blog. For a broader perspective on the capabilities of the Gemini models, you can read our previous analyses on the advances in generative AI and Google's benchmarks.