Managed Agents within the Gemini Interactions API now feature enhanced sandbox oversight and flexible model selection. By integrating environment hooks and new automation tools, developers can now audit tool calls in real-time and manage costs through budget controls and a free tier.
Model Selection and Defaults
The antigravity-preview-05-2026 agent has transitioned to Gemini 3.6 Flash as its default model, requiring no code modifications for implementation. Developers may still manually specify models via the agent_config.model parameter during interaction creation. Options include the balanced Gemini 3.6 Flash for coding and reasoning, the previous generation Gemini 3.5 Flash for general workflows, or Gemini 3.5 Flash-Lite to reduce latency and expenses.
Managed Agents in Gemini API now default to Gemini 3.6 Flash. New environment hooks let you block, lint, or audit tool calls inside the sandbox.
Also, we’ve added budget controls, scheduled triggers, and free tier access. Your browser does not support the audio element.
Managed Agents in Gemini API are getting environment hooks, model selection, and free tier access.
Sandbox Environment Hooks
New environment hooks enable the execution of custom scripts before or after tool calls within the isolated cloud sandbox. By utilizing a .agents/hooks.json file, operators can implement pre_tool_execution and post_tool_execution handlers. These hooks support regular expression matching to target specific tools or all actions. Scripts returning a deny decision will cause the runtime to skip the tool call and provide the rejection reason to the model's context.
These capabilities build on our previous release introducing background tasks and remote MCP server integration. With managed agents in the Gemini Interactions API, a single API call coordinates, reasoning, code execution, package installation, file management, and web retrieval inside an isolated cloud sandbox.
If you're using an AI coding assistant, drop this in your terminal to give it access to the Interactions API skill: npx skills add google-gemini/gemini-skills --skill gemini-interactions-api. Below are examples using the @google/genai TypeScript/JavaScript SDK.
For Python or cURL, check out the Antigravity agent documentation.
Validation and Compliance
The update introduces both command-based scripts and HTTP handlers that POST to external endpoints, allowing for production-grade validation pipelines. For instance, investment bank Offdeal employs post_tool_execution hooks to automate image verification for company logos within the remote sandbox. This process involves pixel-level quality checks and Gemini vision verification to ensure images meet specific size, contrast, and transparency requirements before being approved for use.
The antigravity-preview-05-2026 agent now runs Gemini 3.6 Flash by default. Your next interaction picks it up automatically.
You can also explicitly select models by passing agent_config.model when creating an interaction or managed agent. Use Gemini 3.5 Flash-Lite for lower cost, or pin to your model of preference.
Environment hooks let you run your custom scripts before or after every tool call the agent makes inside its sandbox.
Operational Infrastructure
The Gemini Interactions API streamlines complex workflows by coordinating reasoning, package installation, file management, web retrieval, and code execution in a single call. To support broader adoption and operational scaling, Google has introduced scheduled triggers, budget controls, and free tier access. These additions complement previous updates that integrated remote MCP servers and background task capabilities.
Add a .agents/hooks.json into your environment and the runtime executes your handlers on pre_tool_execution or post_tool_execution events. The matcher field supports regular expressions, allowing you to target multiple tools with | or catch everything with *: For complete HTTP hook definitions and failure handling semantics, refer to the hooks documentation.
Teams are already using hooks to build production-grade validation pipelines. For example, AI-native investment bank Offdeal uses post_tool_execution hooks to run automated image verification inside the remote sandbox.
"OffDeal is an AI-native investment bank, and Archie is the AI analyst our bankers use every day.
Key signals
- Gemini 3.6 Flash is now the default for managed agents.
- Environment hooks allow blocking or auditing tool calls via custom scripts.
- Free tier access and budget controls are now available.
- You can also explicitly select models by passing agent_config.model when creating an interaction or managed agent.
- Developers can experiment with agentic workflows using an API key from a project without active billing.
What to watch
Monitor how firms like Offdeal utilize environment hooks to create automated, high-precision validation pipelines within remote sandboxes to replace manual quality assurance for complex assets.
Source and methodology
This Intelligence Daily briefing preserves the key facts published by Google AI and organizes them into a fuller, reader-friendly report. Read the original reporting.