Interactions API: Google makes Gemini default interface (GA)

Google DeepMind has made the Interactions API the default interface for Gemini models and agents, promoting it from beta to general availability on Jun 22, 2026. The Interactions API replaces the older generateContent interface in Google AI Studio and all documentation, though the old API will continue to work.

What changed and when?

The Interactions API left beta in December 2025 and became generally available on Jun 22, 2026, at which point Google set it as the default interface for Gemini models and agents. Google also published a migration guide for developers to move from the legacy generateContent interface to Interactions.

The switch changes both the surface area and the release path: new agent features will only ship through the Interactions API going forward, even as the previous generateContent endpoint remains functional for existing integrations.

How does the Interactions API work compared with generateContent?

Interactions replaces the older role-based message schema with a typed-steps model, where every action from user input to function calls is represented as a defined step. The new schema simplifies interactions by converting the previous role labels such as "user" and "model" into explicit step types, and it offers two operating modes developers can choose between.

Flex mode is billed as the cost-saving option, cutting costs by 50 percent, while Priority mode optimizes for speed. The Interactions API also adds higher-level agent capabilities, including Managed Agents with their own Linux sandbox, background execution for long-running tasks, tool chaining with Google Search and Maps, and media generation for images, music, and speech.

Why did Google make this change?

Google positions Interactions as the platform for new agent experiences: "Interactions sets the stage for the new era of Agents," writes Logan Kilpatrick, Google's developer relations lead. Moving to a typed-steps model standardizes how inputs, outputs, tool calls, and background tasks are represented, which makes it easier to deliver richer agent features such as sandboxed Managed Agents and multi-tool chaining.

Consolidating development and documentation onto Interactions also lets Google route all future agent innovations through a single interface rather than maintaining parity across multiple endpoints. That should speed the rollout of features that depend on the newer schema and execution model.

What to watch

Watch for adoption friction points in the migration guide and for which agent features require the Interactions API exclusively, since Google says new agent capabilities will only ship through Interactions. Also track how quickly third-party tools and integrations switch from generateContent to the typed-steps schema, and whether the Flex mode's 50 percent cost claim changes developer choices between cost and latency.