Google bundles Gemini 3.1 Pro, Deep Think, and creator tools in February app drop

What Google shipped in the February Gemini drop

Google’s February 2026 Gemini app update bundles several model and creator features into one release. The company said Gemini 3.1 Pro and Deep Think are now available in the Gemini app, while Gemini 3.1 Nano Banana 2 becomes available with Pro plans as Google’s latest multimodal image model.

The release also expands creative workflows. Google says Veo Templates add more structured control for video generation, including text, scene, style, and audio directions. In Canvas, users can now generate 30-second songs with Lyria 3 from text and can also turn images into music. Google also said AI Ultra users in the United States can model their own voice inside Canvas.

Why this drop matters

The significance is less any single feature than the packaging. Google is turning the Gemini app into a distribution layer for frontier models, reasoning modes, image generation, music generation, and creator tooling at the same time. That matters because the market is moving away from standalone model launches and toward integrated product surfaces where users compare full suites, not only raw benchmarks.

Google also highlighted workflow changes aimed at making Gemini more persistent inside everyday tasks. Gemini Live can now view both a user’s screen and camera at the same time, giving the assistant more context during multi-step interactions. Scientific citations are also being integrated more deeply into Search and Gemini outputs, which suggests Google is still trying to make trust and traceability visible product features.

What it says about the next phase of consumer AI

The broader pattern is that flagship AI apps are becoming multimodal workspaces rather than single-purpose chat interfaces. Google is using the monthly drop format to ship faster and to make the Gemini app feel like a rolling platform for new models and media tools. That lowers the distribution cost of each new capability while raising expectations that a top-tier assistant should handle reasoning, live context, image work, music, and video from the same front end.

For users and competing vendors, that changes the comparison set. The question is no longer only which company has the best model on a benchmark. It is which company can keep turning frontier research into coherent product workflows fast enough that people build habits around them.

Sources: Google