Run Gemma 4 Locally on Windows and Set Up Code Autocomplete in IntelliJ IDEA

Google released Gemma 4 on April 2, 2026. This guide walks you through the complete setup — autocomplete suggestions appearing inline in your Java editor.

1. Why Gemma 4 for local development?

Gemma 4 is Google DeepMind's latest family of open models, designed as highly capable reasoners with configurable thinking modes, native function-calling support, and notable improvements in coding benchmarks. It ships under an Apache 2.0 license — meaning you can run it commercially with no usage fees and no code leaving your machine.

For Java developers on IntelliJ, the value proposition is simple: GitHub Copilot sends your code to Microsoft's servers. Gemma 4 running on your own hardware does not. And you don't need to have a GitHub Copilot subscription.

2. Hardware requirements — pick your model variant

I have a medium hardware on windows machine:

• 16GB RAM
• RTX 3060 - 12GB VRAM

So I will use google/gemma-4-e4b, this model requires only 6GB RAM or VRAM. You can get the model from https://deepmind.google/models/gemma/gemma-4/ or from LMStudio: https://lmstudio.ai/models/gemma-4

When a more complex task is executed GPU is involved, and the consumed resources are as follows:

3. Install LMStudio

To make it simple, let's use https://lmstudio.ai/.

1. Download and install it on your machine.
2. From https://lmstudio.ai/models/gemma-4, click the get button for google/gemma-4-e4b
3. Go to the developer tab in LMStudio and start it with the model

You should have something like this:

3.1. Install the Continue plugin in IntelliJ IDEA

Continue is the open-source plugin that bridges your IDE to LMStudio. It supports code completion, converting comments into code, and debugging assistance — and works entirely locally without sending your code to external servers and is free.

Install via the IntelliJ plugin marketplace:

• Open IntelliJ IDEA → File → Settings → Plugins
• Select the Marketplace tab and search for Continue
• Click Install, then restart IntelliJ when prompted
• The Continue panel appears in the right sidebar after restart

Alternatively, install from within the IDE using the keyboard shortcut Ctrl+Alt+S to open Settings, then navigate to Plugins.

4. Configure Continue for Gemma 4 chat + autocomplete

You can do it from IntelliJ IDEA continue plugin:

1. Click the Continue plugin Icon
2. In the continue Chat, click Add chat Model
3. Choose LMStudio from the list and click Connect. A default configuration will be generated
4. 
5. Go to Continue Settings, click Models, and click the settings of the model
6. Update the code with:

name: Local Config
version: 1.0.0
schema: v1
models:
  - name: google/gemma-4-e4b
    provider: lmstudio
    model: google/gemma-4-e4b
    apiBase: http://localhost:1234/v1/
    roles:
      - autocomplete
      - chat
      - edit
      - apply
    defaultCompletionOptions:
      temperature: 0.2
      contextLength: 32768
    autocompleteOptions:
      debounceDelay: 100
      maxPromptTokens: 2024
      multilineCompletions: auto

Or update the Continue configuration from a config.yaml file located at C:\Users\<your-name>\.continue\config.yaml on Windows.

Open this file in any text editor and replace its contents with the following configuration. It sets up two separate model roles: Gemma 4 for chat/editing tasks and autocomplete.

name: Local Config
version: 1.0.0
schema: v1
models:
  - name: google/gemma-4-e4b
    provider: lmstudio
    model: google/gemma-4-e4b
    apiBase: http://localhost:1234/v1/
    roles:
      - autocomplete
      - chat
      - edit
      - apply
    defaultCompletionOptions:
      temperature: 0.2
      contextLength: 32768
    autocompleteOptions:
      debounceDelay: 100
      maxPromptTokens: 2024
      multilineCompletions: auto

5. Enable and test TAB autocomplete in IntelliJ

Enable the setting

In IntelliJ, navigate to File → Settings → Tools → Continue and check the box for Enable Tab Autocomplete.

Test it in a Java file