NotebookLM gets stronger with Gemini 3: the Data Tables feature has arrived

NotebookLM, Google's AI tool focused on note-taking and research, has long done a good job of “compile, summarize and connect from your sources”. With the new update, it has moved to a more powerful model under the hood: NotebookLM now runs on Gemini 3. The Data Tables feature arriving at the same time takes the output into a more “workable” format.

In this post, I'll clarify two things:

• What does the move to Gemini 3 improve in NotebookLM in practice?
• What problems does the Data Tables output solve, and how do you use it?

> Note: Access to features in NotebookLM can change over time. Some new features may first open to paid plans and then gradually roll out to free users.

What does NotebookLM's move to Gemini 3 mean?

When the underlying model changes, you may not see a big “button” in the user interface; most of the time the difference shows up in daily use. The gains expected in NotebookLM with the move to Gemini 3 are:

• Better reasoning: Sorting out contradictions across multiple sources, establishing cause-and-effect relationships, drawing inferences.
• Stronger multimodal understanding: An experience that gets closer to better understanding content types beyond text (depending on the use case and the source you upload).
• More consistent synthesis: Producing more “balanced” answers within a similar frame even when you ask the same question in a different way.

In short, NotebookLM is moving from being “an assistant that answers quickly from your notes” closer to the role of a research and planning assistant.

The new Data Tables: what does it change in NotebookLM?

NotebookLM's strength was reading sources and making sense of them; its weakness was that the resulting text wasn't always “operational”. This is exactly where the Data Tables output comes in:

• It moves scattered text into a column/row layout
• It gathers the same kind of information into a single format
• It saves time on tasks like comparison, tracking and exporting

This feature shows its impact most clearly in work that involves “many sources + many criteria”.

In which scenarios is it really useful?

The examples below show where the Data Tables output is strongest:

• Turning meeting notes into an action plan: A tracking table with columns like “Task”, “Owner”, “Priority”, “Date” and “Dependencies”.
• Comparing research summaries: Comparing several papers with columns like “year”, “sample”, “finding” and “limitations”.
• Product/tool evaluation: Seeing alternatives in a single table by criteria like “price”, “features”, “pros/cons” and “compatibility”.
• Education/learning: Studying historical events by splitting them into fields like “period”, “actors”, “cause” and “outcome”.
• Daily planning: Columns like “destination”, “ideal date”, “estimated cost” and “notes” for a vacation/route plan.

What these examples have in common: once you turn text into a table, taking the next step gets easier.

Exporting: Sheets and the document flow

The most critical advantage of the Data Tables output is that the table is “portable”. For many people, the real work doesn't happen in NotebookLM but afterwards:

• Filtering/sorting in Google Sheets
• Sharing with the team
• Moving the table into a project management tool

That's why the table output brings NotebookLM closer to being a tool that “turns things into work” rather than one that “reads and explains”.

Who should try it right away?

• Researchers: Those who want to compare output from multiple sources in an organized way
• Product/operations teams: Those who want to turn meeting notes into action
• Students and content creators: Those who want to keep complex information in an orderly framework

Final word

The move to Gemini 3 strengthens NotebookLM's “understanding and inference” side, while the Data Tables feature ties that power to a more actionable output. If you use NotebookLM regularly, this update can seriously speed up your daily flow, especially for planning and comparison work.