The Gemini app is receiving a substantial update this February, rolling out recent features designed to enhance both creative expression and complex problem-solving. Google’s latest “Gemini Drop” introduces Lyria 3, a high-fidelity AI music generator, alongside significant improvements to the core Gemini 3.1 model and tools for visual content creation. These updates aim to position Gemini as a more versatile and integrated AI partner for a wider range of users, from casual creators to researchers tackling intricate challenges. The focus is on empowering users with AI capabilities that feel intuitive and accessible, while also respecting artistic integrity and intellectual property.
At the heart of this update is Lyria 3, Google’s most advanced AI music model to date. This new tool allows users to generate 30-second musical tracks based on simple text or image prompts. Imagine turning a funny moment or a cherished photo into a custom soundtrack complete with instrumentals, vocals, and lyrics. The possibilities are broad, ranging from creating background music for focused function to crafting shareable anthems for social media. According to Google, Lyria 3 is designed to be respectful of copyright, creating tracks that are *inspired* by artists rather than directly imitating them. The service also incorporates an imperceptible “watermark” to identify the music as AI-generated, addressing concerns about authenticity and attribution. This feature is currently available in beta in all countries where the Gemini app is accessible, with a minimum age requirement of 18.
Gemini 3.1 Pro: A Leap in Reasoning Capabilities
Beyond music creation, Google is significantly boosting Gemini’s core intelligence with the release of Gemini 3.1 Pro. This upgrade is specifically tailored for demanding workflows that require advanced reasoning skills. The model has demonstrated a substantial improvement in logic capabilities, achieving a verified score of 77.1% on the ARC-AGI-2 benchmark – more than double the performance of its predecessor, Gemini 3 Pro. This enhanced intelligence translates to practical applications for university students, researchers, and developers. Gemini 3.1 Pro excels at synthesizing data, building interactive designs, and even generating website-ready animated SVGs directly from text prompts. It can also handle “creative coding,” adapting the style of an image or document to inform the design of a website interface.
For users requiring even more specialized reasoning power, Google is offering “Deep Think,” a specialized mode within Gemini 3.1. However, access to Deep Think is currently limited to subscribers of Google AI Ultra. Both Gemini 3.1 Pro and Deep Think are available to developers through the Gemini API, accessible via Google AI Studio, Gemini CLI, Vertex AI, and Antigravity, according to Case Western Reserve University’s UTech news.
Enhanced Visual Creation with Nano Banana 2 and Veo Templates
The Gemini update isn’t solely focused on audio and reasoning. Visual content creation is also receiving a boost with the introduction of Nano Banana 2, Google’s newest image model. Nano Banana 2 allows users to add text in any language to their images with improved accuracy and speed. This feature is particularly useful for creating visually engaging content for diverse audiences. Complementing Nano Banana 2 are Veo Templates, which provide a starting point for video creation. Users can browse a gallery of pre-designed styles, remix them with their own details, and produce polished, unique videos.
Streamlined Research with Verified Scientific Citations
Recognizing the importance of reliable information, Google is integrating verified scientific citations directly into Gemini’s search results. This feature provides users with direct links to scientific papers, eliminating the need to sift through countless search pages to find credible sources. This is a significant step towards ensuring that research is backed by high-quality data and promoting academic integrity. The ability to quickly access primary sources will be invaluable for students, researchers, and anyone seeking evidence-based information.
These updates, collectively known as the Gemini Drop, represent a significant evolution for Google’s AI platform. The combination of creative tools like Lyria 3 and Nano Banana 2, alongside the enhanced reasoning capabilities of Gemini 3.1 Pro, positions the app as a powerful and versatile assistant for a wide range of tasks. The emphasis on accessibility, artistic respect, and verifiable information underscores Google’s commitment to responsible AI development.
Google plans to continue refining and expanding Gemini’s capabilities throughout 2026. The next scheduled update, focusing on improved multilingual support and expanded API access for developers, is expected in late spring. Users can find more information and explore the new features on the Gemini Drops Hub.
What are your thoughts on the new Gemini features? Share your experiences and ideas in the comments below.
