Google has announced the integration of its LearnLM model directly into Gemini 2.5, along with additional updates to its learning products.
Ben Gomes, Chief Technologist for Learning & Sustainability at Google, said: "Making knowledge accessible to everyone has always been our highest priority, which is why people turn to our products each day to help them learn — for school, work or life. With AI, we can do this at a speed and scale never before possible, and make the process of learning more active, engaging and effective. By building tools that enable you to keep pace with your own curiosity, in formats that match your goals and preferences, we hope to help everyone in the world learn anything in the world."
Gomes highlighted the importance of advanced models in supporting learning. "This fundamentally starts with really capable models. We can make models even better when we refine them for specific uses. That's why last year, we introduced LearnLM: our family of models and capabilities fine-tuned for learning. For years we've been working with education experts to research, measure and improve on building AI systems that support effective learning practices."
The company stated that the latest update marks the integration of LearnLM into Gemini 2.5, asserting Gemini 2.5 as the "world's leading model for learning." Google reported that Gemini 2.5 Pro outperformed competitors across categories of learning science principles, and was preferred by educators and pedagogy experts for supporting users' learning goals and adhering to key principles of good pedagogy across various scenarios.
According to Google, the infusion of LearnLM enables Gemini to "go beyond just giving you the answer" by focusing on explanation and reasoning. The company stated: "By applying LearnLM capabilities, and directly incorporating feedback from experts across the industry, Gemini adheres to the principles of learning science to go beyond just giving you the answer. Instead, Gemini can explain how you get there, helping you untangle even the most complex questions and topics so you can learn more effectively. Our new prompting guide provides sample instructions to see this in action."
Google pointed to research showing that multimodal information—presented via audio, video, images, and text—improves learning outcomes. With Gemini's multimodal functionality, users will have greater control over how information is consumed, tailored to personal needs and preferences.
The company highlighted updates to NotebookLM, a tool designed for studying and research using a custom set of sources. Features such as Audio Overviews, now available in more than 80 languages, and Mind Maps provide different pathways for content exploration. Google indicated that more flexibility is being introduced to Audio Overviews, enabling users to select summary lengths. This feature will roll out in English first and expand to additional languages.
Additionally, Google is developing Video Overviews, which will enable users to transform notebook content into educational videos. The company noted that user feedback indicated a desire for more visual content during overviews, prompting this forthcoming enhancement.
For Google Search, the company stated that users increasingly rely on it for learning, with AI Mode now providing advanced reasoning and multimodality for deeper exploration, including links to web sources and follow-up question functionality. Google shared plans to introduce Deep Search to AI Mode and announced that AI Mode is now available to all users in the United States.
Google is also implementing a custom version of Gemini 2.5 within both AI Mode and AI Overviews in the United States for delivering more detailed responses supported with web links. The company announced: "Learning also takes place in the context of the world around us. So we're taking a big step in multimodality by bringing Project Astra's live capabilities into AI Mode with Search Live. Beyond asking questions with text and images, soon you'll be able to show Search what you see and ask questions about things in the world around you in real-time. Search will provide helpful information with links to explore along the way as you go back-and-forth. Search Live is coming to Labs this summer, perfect for heading back-to-school."
The Gemini app will now provide a free AI Pro upgrade to students in the United States, Brazil, Indonesia, Japan, and the United Kingdom who sign up by 30 June 2025. Eligible students receive 15 months of free access to Google AI Pro, including 2 TB of storage and access to NotebookLM, to assist with writing, studying, and homework.
Google also announced global availability, starting immediately, for custom quiz creation through Gemini for students aged 18 and over. This feature allows students to generate interactive practice quizzes on any topic or uploaded materials, with the system providing hints, explanations for correct and incorrect answers, and summarising areas of strength and those needing further study.
Ongoing experiments include Sparkify, an upcoming feature to create short animated videos from user questions or ideas using Gemini and Veo models, and a conversational tutoring prototype in Project Astra. This tutor is designed to guide students through homework, offering step-by-step problem-solving assistance and generating explanatory diagrams when necessary.
Google is updating its Learn About project as well. The company stated: "We're also bringing improvements based on your feedback to Learn About, an experimental Labs project where conversational AI meets your curiosity. Through LearnLM capabilities now in Gemini models, we can deliver even more nuanced explanations and relevant connections. We're making this experience available for more learners (including teens), adding session history so you can pick up where you left off and offering the ability to upload your own source documents so Learn About can ground explanations in your course materials, notes or research papers."
