Google adds hooks, budget controls, and free tier to Managed Agents in Gemini API

Managed Agents in the Gemini API now default to the Gemini 3.6 Flash model, replacing the previous unspecified default. No code changes are required for existing users to adopt the new default during their next interaction.

Developers can explicitly select models by passing the agent_config.model parameter when creating an interaction or managed agent. Supported models include Gemini 3.6 Flash (default), Gemini 3.5 Flash, and Gemini 3.5 Flash-Lite. The latter is positioned as the lowest-latency and lowest-cost option within the Gemini 3.5 family.

New environment hooks let teams run custom scripts before or after tool calls inside the sandbox. Hooks can block, lint, or audit tool calls using pre_tool_execution and post_tool_execution events. Hooks support regular expression matchers to target specific tools or all tools (e.g., code_execution, write_file).

Hooks can execute local commands or POST to external endpoints via HTTP handlers. If a pre_tool_execution hook returns a decision to deny, the tool call is skipped and the reason is passed into the model’s context. Teams are already using hooks to build production-grade validation pipelines, such as automated image verification for AI-native workflows.

Additional features include budget controls, scheduled triggers, and free tier access for Managed Agents. These capabilities build on prior releases that introduced background tasks and remote MCP server integration.