Gemini 3.5 and Antigravity come to Google NotebookLM
Google’s NotebookLM is receiving a major update, integrating the latest Gemini 3.5 Flash model and Antigravity. This enhances its analytical capabilities, allowing for faster processing, improved accuracy, and code execution for research. The update also introduces new output formats and streamlined web source integration for a more comprehensive user experience.
Google’s NotebookLM is receiving its most significant update yet, incorporating the advanced Gemini 3.5 Flash model and Antigravity technology. This upgrade promises faster, more efficient processing and improved analytical capabilities for users. Gemini 3.5 Flash, which debuted at Google I/O, is designed to enhance performance and reduce token costs, benefits that are now extending to NotebookLM.
NotebookLM, launched in 2023, leverages Google's AI models to analyze user-provided sources like documents and webpages. With the new Gemini 3.5 integration, Google reports a 65% win rate in internal evaluations compared to the previous model, demonstrating significant improvements across categories like Accuracy, Multilingual Support, and Large Document Analysis.
The inclusion of Antigravity enables NotebookLM to write and execute code for research purposes, significantly expanding its utility. Google states that NotebookLM will offer over 100 software skills to streamline workflows, allowing users to achieve tasks that previously required multiple applications.
Beyond enhanced analytical capabilities, NotebookLM will now support diverse output formats. The research bot can generate documents, infographics, quizzes, and audio overviews, which are managed within the new Studio Panel. Users can also prompt NotebookLM to edit these files post-creation, with more file types expected in the future.
To further streamline research, NotebookLM has improved its web source integration. Users can now ask Gemini directly within the chat interface to find relevant web sources, which are then presented in a "research report" for easy import. This ensures that all future interactions with the notebook will utilize both user-provided and AI-discovered sources.
These new features are rolling out gradually. Initially, they will be available to AI Ultra subscribers and Workspace business customers with AI Ultra Access and AI Expanded Access, with broader availability to other Google accounts expected soon.
Related articles
Build real agentic apps using CUGA: two dozen working examples on a lightweight harness
CUGA, IBM's open-source Agent Harness, simplifies building agentic applications by handling infrastructure, allowing developers to focus on tools and prompts. It offers pre-assembled components for planning, execution, and state management, significantly reducing development time. CUGA has topped agent benchmarks like AppWorld and WebArena.
OpenAI launches new initiative to help find and patch open source bugs
OpenAI has launched "Patch the Planet," a new initiative in partnership with cybersecurity firm Trail of Bits, to enhance the security of open-source projects. This program aims to assist maintainers in identifying and patching bugs, utilizing OpenAI's AI-powered security tools while reducing the burden on project teams.
PP-OCRv6 on Hugging Face: 50-Language OCR from 1.5M to 34.5M Parameters
Baidu has released PP-OCRv6, an advanced optical character recognition (OCR) model supporting 50 languages. Available on Hugging Face, this version significantly improves accuracy and efficiency across various parameter sizes, from 1.5 million to 34.5 million, marking a substantial leap in multilingual OCR technology.