Google Upgrades NotebookLM With Gemini 3.5 Flash and Antigravity Cloud Computer
Summary
- • NotebookLM upgraded to Gemini 3.5 Flash, debuted at Google I/O 2026, bringing faster and more cost-efficient AI processing.
- • Google claims 65% win rate over Gemini 3.1 across five core evaluation dimensions including accuracy and multilingual support.
- • Antigravity integration gives NotebookLM an embedded cloud computer to write and execute code within research workflows.
- • Over 100 built-in software skills enable multi-app workflow automation directly inside notebooks.
Details
Gemini 3.5 Flash upgrade
NotebookLM moves from Gemini 3.1 to Gemini 3.5 Flash, promising faster processing and lower token costs while delivering equal or better output quality.
65% win rate vs old model
Google internal evaluations: Gemini 3.5-powered NotebookLM wins 65% of comparisons across Accuracy & Quality, Multilingual Support, Large Document Analysis, Document Creation, and Advanced Research.
Antigravity cloud computer
Embedded cloud compute environment allows NotebookLM to write and run code in service of research goals — transforming it from a Q&A interface into an active computational agent.
100+ software skills
Built-in skill library enables notebook-based automation of workflows that previously required switching between multiple apps.
NotebookLM history
Launched in 2023 at the start of the AI boom, NotebookLM has outlasted many cancelled Google AI initiatives and is now receiving its most significant upgrade to date.
Rows cover the model upgrade, performance benchmarks, Antigravity compute feature, skill library, and product history.
What This Means
The Gemini 3.5 + Antigravity upgrade marks a meaningful transition for NotebookLM — from a smart document Q&A tool into an active research agent capable of writing and running code against your sources. The 65% win rate claim, while methodologically vague, signals that Gemini 3.5 Flash improvements are now reaching end users in production products. Adding code execution is a competitive differentiator that positions NotebookLM against more advanced analysis platforms. For researchers, knowledge workers, and enterprises, this update makes NotebookLM a more powerful hub for document-grounded AI workflows.