Google makes Interactions API the default interface for Gemini models and agents
Quick Answer
Google DeepMind has transitioned to the Interactions API as the default interface for its Gemini models and agents, replacing the previous generateContent API.
Quick Take
This new API features a simplified schema with typed steps, and future agent functionalities will exclusively utilize this interface.
Key Points
- The Interactions API simplifies the interface with typed steps instead of role-based structures.
- Future agent features will exclusively launch through the Interactions API.
- This change marks a significant shift in how Gemini models will operate.
📖 Reader Mode
~1 min readGoogle DeepMind has released the Interactions API as the default interface for Gemini models and agents. The API, in beta since December 2025, is now generally available and replaces the old generateContent interface in Google AI Studio and all documentation. The old API still works, but new agent features will only ship through the Interactions API going forward. "Interactions sets the stage for the new era of Agents," Google's developer relations lead Logan Kilpatrick writes. Google has published a migration guide for the switch.
Recent additions include 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.
Google also simplified the schema since the beta. The old role structure with labels like "user" and "model" is now replaced by typed steps, where every action from user input to function calls is its own defined step. Developers can choose between Flex and Priority mode. Flex cuts costs by 50 percent, while Priority optimizes for speed.
— Originally published at the-decoder.com
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from The Decoder
See more →
An AI model programmed nonstop for 19 days on a single MirrorCode task that cost $2,600 to run
Epoch AI's MirrorCode benchmark reveals Claude Opus 4.7 as the leader with a 56% solve rate, reconstructing a 16,000-line toolkit in 14 hours. Despite this, all models tested struggle with the most complex tasks, highlighting limitations in current AI capabilities. The single task consumed $2,600 over 19 days, raising questions about cost-effectiveness in AI development.
