Google’s NotebookLM, one of the company’s first generative‑AI experiments, survives the tides that swept away other projects. Today it receives its biggest upgrade yet, shifting to Gemini 3.5, adding new file‑type support, and tightening web‑source integration. The move promises faster, cheaper responses while keeping the same or higher output quality for developers and data scientists alike.

Gemini 3.5 Flash debuted at Google I/O, boasting faster processing and lower token costs. Google claims businesses can cut expenses while matching or exceeding earlier model quality. These gains now filter into NotebookLM, a 2023 launch that lets users analyze documents and webpages with Google’s latest AI engines and support advanced research workflows for teams worldwide.

Side‑by‑side tests compare the old Gemini 3.1 branch with the new 3.5. Google reports NotebookLM wins 65 percent of the time across five core dimensions—accuracy, multilingual support, large‑document analysis, document creation, and advanced research. The upgraded model also gains a dedicated “cloud computer” that runs Antigravity code to automate data pipelines and streamline code execution for analysts.

NotebookLM now hosts over 100 software skills, letting users build notebook workflows without jumping between apps. The integration of Antigravity expands coding capabilities, while the Flash model trims token costs. With these enhancements, Google positions NotebookLM as a versatile research companion, ready for enterprises that demand efficient, multilingual AI analysis across multiple industries today.