NotebookLM gets Gemini 3.5 and Antigravity for deeper queries
NotebookLM
Gemini
Google
TLDR: SAN FRANCISCOāGoogle NotebookLM is upgrading to Gemini 3.5, adding more file types and smoother web source integration, plus Antigravity to process more from queries.
Key Takeaways:
- NotebookLM started in 2023 as a source driven AI tool for documents and webpages, and it is still active while other early bets aged out.
- The update brings Gemini 3.5 model support, more file formats, streamlined web source integration, and embedded Antigravity for deeper query handling.
- Google reports NotebookLM averaged a 65 percent win rate versus an older Gemini 3.1 branch across Accuracy and Quality, multilingual support, and large document analysis.
Google is using NotebookLM like a runway for its newest model upgrades, and Antigravity hints it wants fewer āask againā moments from busy researchers.
Google is using NotebookLM like a runway for its newest model upgrades, and Antigravity hints it wants fewer āask againā moments from busy researchers.
Q&A
What will Antigravity change about how people structure prompts inside NotebookLM?
It is designed to let NotebookLM extract more value from each query, so users may rely less on multi prompt workflows and more on tighter questions tied to sources.
Why does moving to Gemini 3.5 matter more than simply faster responses?
Gemini 3.5 improvements also target accuracy, multilingual support, and large document handling, which affects whether summaries stay trustworthy when documents get messy.
Could the new file type support shift NotebookLM from note assistant to light production tool?
More supported inputs reduce friction for real work, so people can draft and iterate faster without converting everything into a single accepted format.
What does the 65 percent win rate suggest about Googleās internal evaluation bar?
It signals Google believes the 3.5 upgrade clears multiple benchmarks at once, not just one metric, and that it will likely push similar upgrades into other Workspace AI features.
If companies worry about token costs, why does Google keep rolling Flash style gains into other products?
Cost control is a distribution strategy: lower processing expense helps Google scale AI usage, making it easier to keep NotebookLM broadly usable without throttling.
No comments yet. Be the first to share your thoughts!