Google Gemini February Drop Adds AI Music Creation and Enhanced Reasoning

Google rolled out its February 2026 Gemini Drop on Thursday, packing the AI assistant with music generation capabilities, upgraded reasoning models, and faster image creation tools.

The headline feature is Lyria 3, Google’s most advanced music model to date. Now in beta, it lets users generate 30-second custom soundtracks from text prompts or images. Want background music that matches a photo’s mood? Describe it or upload the image, and Lyria handles the rest.

On the intelligence front, Gemini 3.1 arrives with what Google calls “significantly improved” problem-solving for complex workflows. The release splits into two tiers: 3.1 Pro for demanding tasks, and Deep Think—a specialized reasoning mode targeting scientific and engineering applications. That last one’s locked behind Google AI Ultra subscriptions.

The update lands amid a busy week for Google’s AI efforts. Just two days prior, the company announced multi-step task automation for Galaxy S26 and Pixel 10 devices, letting Gemini execute sequences of actions across different Android apps. That feature builds on the newly detailed “AppFunctions” platform, which essentially gives Gemini hooks into third-party applications.

Visual and Research Tools Get Upgrades

Nano Banana 2, Google’s updated image model, promises higher fidelity outputs with one practical addition: text rendering in any language with “real-world accuracy.” Anyone who’s fought with AI image generators to produce readable text knows that’s been a persistent weak spot across the industry.

Veo Templates offer a different angle on content creation. Rather than starting from scratch, users can browse a gallery of video styles, pick one, and remix it with custom details. It’s the kind of guided approach that lowers the barrier for users who know what they want but not how to prompt for it.

For researchers, Gemini now surfaces verified scientific citations with direct links to source papers. Google’s framing this as a time-saver—no more digging through search results to find the actual study behind a claim.

Context and Competition

These updates continue Google’s aggressive repositioning of Gemini since rebranding from Bard in February 2024. The assistant now spans mobile, web, and increasingly integrates with Workspace apps like Gmail and Calendar. The enterprise angle got a boost this week too, with Google releasing dedicated Gemini Enterprise mobile apps aimed at workplace productivity.

The Deep Think reasoning mode puts Google in direct competition with OpenAI’s o1 and similar “thinking” models from Anthropic. Whether the Ultra subscription paywall limits adoption remains to be seen, but it signals where Google sees premium value in AI capabilities.

Full details on the February Drop are available through Google’s Gemini Drops Hub.

Image source: Shutterstock