Netlify Adds Gemini 3.8 Flash support Across Functions and Agent Runners.

Netlify has made Google’s Gemini 3.8 Flash model available across every type of Netlify Function and inside Agent Runners. For developers already using the Google GenAI SDK, adopting it can be as simple as replacing the model name with gemini-3.8-flash.

The deployment platform announced the immediate availability in a LinkedIn post and published sample code in its official changelog. Netlify says no new API key or additional authentication code is required when switching an existing Google GenAI integration.

That small implementation detail is the important part. Model upgrades can otherwise involve changing providers, secrets, request formats or infrastructure. Netlify is positioning its AI Gateway as the stable layer underneath those changes, allowing application code to continue using the provider’s standard SDK.

Netlify’s example imports the Google GenAI SDK inside a Function, creates a client without supplying credentials and identifies the new model in the content-generation request:

import { GoogleGenAI } from '@google/genai';

export default async () => {
  const ai = new GoogleGenAI({});
  const response = await ai.models.generateContent({
    model: 'gemini-3.8-flash',
    contents: 'How can AI improve my coding?'
  });

  return Response.json(response);
};

For an existing Function that already calls Gemini through the same SDK, the relevant migration is the model value. The rest of the Function can remain in place, according to Netlify. New implementations will still need the SDK and application logic shown in the example, so “zero configuration” applies primarily to Netlify’s authentication and gateway infrastructure rather than the entire application.

AI Gateway sits between the deployed application and the model provider. It automatically handles three operational concerns cited by Netlify:

• Authentication, removing the need to place a Google API key directly in the Function
• Rate limiting, providing controls around requests passing through the gateway
• Caching, which can avoid repeating eligible model calls

The model works with all Netlify Function types, not just a particular runtime pattern. It is also available to Agent Runners, Netlify’s system for assigning software-development work to coding agents.

Support inside Agent Runners gives teams another model option for agent-driven development workflows without requiring separate credential management. That expands on Netlify’s recent work around agent control, including clarifying questions that let Agent Runners resolve ambiguity before building and Ask mode for investigating an application before making changes.

For small development teams, the broader value is portability at the configuration level. A team can test or adopt a newly supported Gemini model without redesigning how secrets reach its serverless code. The application continues to call the Google SDK, while Netlify manages the platform layer behind that call.

There are still unanswered commercial details. The announcement does not specify whether access requires a particular Netlify plan, how model usage is billed or what limits apply. Developers evaluating the model for production should consult the AI Gateway documentation and Agent Runners documentation for the account and operational requirements relevant to their projects.

What is Netlify Gemini 3.8 Flash support?

It is Netlify’s integration of Google’s Gemini 3.8 Flash model with AI Gateway, all Netlify Function types and Agent Runners. Developers can call the model through the Google GenAI SDK while Netlify handles authentication, caching and rate limiting.

How does an existing Function switch to Gemini 3.8 Flash?

A Function already using Gemini through the Google GenAI SDK can set its model string to gemini-3.8-flash. Netlify says the switch does not require a new API key or additional authentication code.

When is Gemini 3.8 Flash available on Netlify?

Netlify announced immediate availability on September 2, 2026. The company says it is available across every Function type and within Agent Runners.