Google Updates Gemini Web and Mobile Interface to Separate Stable and Experimental Tools

In early February 2026, Google began rolling out a structural update to the Gemini web interface and mobile apps. The change reorganizes how features are presented within the service, creating a clear distinction between established tools and experimental capabilities to better manage user expectations regarding stability and performance.

Experimental Features (Labs)

A new Experimental features section, identified by a Labs badge and a beaker icon, has been introduced to host capabilities that are currently under active refinement. These features are segmented by availability:

• Agent (Exclusive to AI Ultra): A key capability in the Labs lineup, this tool allows Gemini to handle multi-step task completion. It enables the assistant to plan and execute workflows that go beyond simple text generation, reflecting the continued development of agentic automation.
• Personal Intelligence (Available to Paid Subscribers): Accessible to both AI Plus and AI Ultra members, this feature is designed to leverage context from past interactions. It helps Gemini adapt to user preferences and working styles over time, offering more relevant responses.
• Visual Layout / Dynamic View (Available to All Users): This interface option is open to free and paid users. It offers a flexible way to view responses, allowing the workspace to adapt dynamically—useful for comparing data or reviewing complex outputs alongside the chat window.

Tools Section

The primary Tools menu remains the home for functionalities that have graduated from the experimental phase. It includes widely accessible tools like Deep Research, Create images, and Canvas. It also integrates premium capabilities for subscribers, such as Create videos (powered by Veo 3.1) for AI Plus members and the reasoning model Deep Think for AI Ultra members, placing them alongside standard tools rather than in a separate testing environment.