Google's Gemini app tests 'Thinking level' with deeper AI customization
Google has begun testing a new 'Thinking level' feature in its Gemini app. The option is currently available to a limited group of users and offers different modes for response depth. It arrives alongside plans to expand third-party app integrations within the platform.
The 'Thinking level' menu appears when users select either the 'Fast' or 'Gemini 3.1 Pro' model. Two settings are available: 'Standard' and 'Extended'. Notably, the feature does not show up when the basic 'Thinking' mode is chosen. For developers, Google AI Studio provides three tiers—'Low', 'Medium', and 'High'—to adjust response complexity.
Gemini is also broadening its connections with external services. Existing integrations include GitHub, OpenStax, Spotify, and WhatsApp. Soon, users will see Canva, Instacart, and OpenTable added to the list. With Canva, they can create designs, manage assets, and collaborate through voice or text prompts. The Instacart integration lets them check grocery stock and add items directly to their shopping cart. More third-party partnerships are expected in the coming months.
The 'Thinking level' rollout remains restricted for now, but its arrival signals Google’s push for more customisable AI interactions. Expanded app integrations will further embed Gemini into daily tasks, from grocery shopping to design work. Users can expect these features to become more widely available over time.
