Google prepares new Ultra mode for AI Studio Build

Google is working on a new Ultra mode for Google AI Studio Build. We spotted the unreleased option in a mode dropdown alongside Plan, Build, and the previously discovered Security mode, which checks apps for vulnerabilities. Ultra’s description says it uses “advanced skills and tools,” but the specific capabilities remain unclear.

In our early tests, generations produced through this flow were quite comprehensive despite using Gemini 3.8 Flash. Based on those results, Ultra appears to devote more thinking effort to app generation than the standard flow. This could help developers tackle more complex builds, although the tests don't show how the mode allocates reasoning or which additional tools it uses.

The feature could fit Google’s broader push toward longer coding workflows. Google has announced Gemini 4 Argon and says it will begin wider access for paid API customers and Google AI Ultra subscribers after an initial testing phase. The announcement provides no firm release date.

Ultra mode could arrive alongside that expansion, but any connection to Argon remains unconfirmed. It is also unclear whether the mode will require an Ultra subscription. Although its description points to advanced skills, there is little reason so far to treat it as equivalent to OpenAI’s Ultrafast mode.

Sources and related context

• Google announces Gemini 4 Argon and its phased rollout: Supports: Google describes Argon's long-running coding workflows and says wider access will start with paid API customers and Google AI Ultra subscribers after initial testing, without specifying a firm release date.
• Google AI Studio Build documentation: Related context: Google explains Build's app-generation workflow and existing agent capabilities, providing background on the platform where we spotted the unreleased Ultra option.