Full Breakdown
Google Unveils Major NotebookLM Upgrade Powered by Gemini 3.5 and Antigravity
6/9/2026, 12:30:41 PM
Upgrade Overview: Gemini 3.5, Antigravity, and Cloud Compute
On June 8, 2026 Google announced that NotebookLM now runs on the Gemini 3.5 large-language model and the Antigravity coding engine. The upgrade adds a secure cloud compute environment that can write and execute code, expands built-in software skills to over 100 curated functions, and supports export to PNG, SVG, PDF, DOCX, Markdown, CSV, JSON, XLSX and PPTX.
Background: NotebookLM’s 2023 Launch and Gemini 3.5 Flash
NotebookLM launched in 2023 as Google’s first generative-AI note-taking tool. Earlier this year Gemini 3.5 Flash debuted at I/O, promising faster processing and lower token usage. Google suggested the model can cut costs while maintaining or improving output quality.
Expanded Capabilities and Reported Performance
The updated NotebookLM can retrieve multilingual primary sources, build a project-specific knowledge base, and show step-by-step reasoning. Over 100 Antigravity-powered software skills enable data cleaning, code generation and visualization within the notebook. Internal tests show an average win rate above 65% across five evaluation dimensions, 69.9% on large-document analysis, and 78.2% on advanced web-research tasks.
Google’s Official Position
Google describes the upgrade as delivering more accurate and reliable information with greater transparency into the model’s reasoning. It highlights the chat interface’s ability to auto-assemble source repositories and provide step-by-step answer explanations, and notes token-cost savings with the Flash model. The features debut for AI Ultra and Workspace business customers.
Criticism and Unresolved Issues
Analysts point out that Google’s win-rate figures are based on internal “top five core evaluation dimensions” that are not publicly defined, preventing independent verification. No third-party benchmarks or details on test data have been released, leaving the robustness of the reported improvements uncertain. Google has not disclosed the size or composition of the datasets used in these internal tests.
Verbatim Quotes
- “even more accurate and reliable information along with better visibility into the thinking process.” — Google spokesperson
- “…the upgraded NotebookLM achieved an average win rate of over 65% — a 15% point margin above parity — across our top five core evaluation dimensions.” — Google
- “Google said that with the latest update, you can start a chat about a project with the app, and it will help you build the knowledge base by suggesting different sources using its research skills and Google Search.” — Google
- “Google has claimed that companies worried about token costs can save big by moving their projects to the new Flash model while also getting outputs that are of similar or better quality.” — Google
Future Rollout
The enhancements are live today for Google AI Ultra users and Workspace business customers with AI Ultra access, with broader availability to additional subscription tiers planned over time.
