How to Use Grok-4 Without Paying for SuperGrok Subscription via LM Arena (Full Guide 2026)

Open your preferred web browser to begin accessing LM Arena. This can be Chrome, Firefox, Safari, Edge, or any modern browser you regularly use. Type "google.com" into your URL address bar or click your browser's home button if Google is set as your homepage.

Once the Google homepage loads, you'll see the search bar in the center of the screen. Ensure your internet connection is stable before proceeding.

Click inside the Google search bar to activate it. Type "LM Arena" into the search field exactly as written. Google's autocomplete feature may suggest "LM Arena AI" or "LMArena.ai" as potential matches.

Select one of these suggestions or complete typing the full phrase yourself. Press Enter on your keyboard or click the search button to execute the search. Google will display a results page with links related to the LM Arena platform within seconds.

Examine the search results carefully to identify the official LM Arena website. The correct link typically appears as the first or second result and displays "lmarena.ai" or "arena.lmsys.org" in the URL. The official site is maintained by LMSYS, a research organization focused on large language model evaluation.

Click on the official LM Arena link to navigate to their homepage. Avoid clicking on advertisements or unofficial third-party sites that may appear in the search results. Your browser will load the LM Arena homepage within a few seconds.

The LM Arena homepage features a clean, research-oriented design with several navigation options. You'll see references to "Chatbot Arena" and various model comparison tools. This platform allows researchers and users to interact with and compare multiple AI models in one centralized location.

Once the LM Arena homepage fully loads, scan the main navigation menu for available options. The "Battle" option is one of the primary features of the platform and should be prominently displayed in the top navigation bar or main content area. This feature is designed to let users compare two AI models side-by-side in a blind test.

Look for a button or link labeled "Battle" — it may also be called "Arena" or "Chatbot Arena" depending on the current site design. The Battle mode is LM Arena's signature feature, allowing users to submit the same prompt to two anonymous models and vote on which response is better.

Click on the "Battle" button to proceed to the next stage. The page will reload and present you with additional options for engaging with AI models on the platform.

On the Battle page, you'll now see multiple interaction options for working with AI models. Look for the "Direct Chat" option, which allows you to interact with a specific model of your choice rather than comparing two anonymous models. This mode is essential because it lets you select Grok-4 specifically.

The Direct Chat button is typically located near the top of the page alongside other mode options like "Side-by-side" or "Arena." Direct Chat mode gives you full control over which model you communicate with, unlike the blind comparison mode where models are hidden until after you vote.

Click on the "Direct Chat" button to activate this mode. The page will update to show a model selection interface and a chat input area. You should now see a dropdown menu or search field where you can specify which AI model you want to use.

After entering Direct Chat mode, locate the model selection dropdown or search field on the page. This is typically positioned near the top of the chat interface and may be labeled "Select Model" or simply show a default model name. Click on this field to activate the model search functionality.

Type "Grok" into the model search field. The platform will begin filtering the available models to show only those matching your search query. LM Arena hosts dozens of different AI models, so using the search function helps you quickly narrow down to the Grok family of models.

As you type, you should see several Grok model versions appear in the filtered results. The search results will update in real-time as you type each letter.

Review the search results to identify the specific Grok-4 model version you need. Look for "Grok-4-0709" in the list of available models — the "0709" designation indicates this is the July 2009 release version. Different version numbers represent different training dates and capability sets, so selecting the correct version is important.

Click on "Grok-4-0709" from the dropdown menu to select it as your active model. After selecting, the model name should appear in the selection field, confirming your choice. The chat interface is now configured to send all your prompts directly to the Grok-4-0709 model.

You may notice the interface updates slightly to indicate which model is active. At this point, you're ready to begin interacting with Grok-4 completely free of charge.

Locate the text input area at the bottom of the Direct Chat interface where you can type your message to Grok-4. This is typically a large text box with a placeholder message like "Type your message here" or "Enter your prompt." Click inside this text area to activate your cursor.

Type the following test prompt exactly as written: "What model are you?" This simple question serves as a verification test to confirm you're actually communicating with Grok-4. The model should respond by identifying itself, which validates that your connection is working properly.

This test prompt is straightforward but effective for confirmation purposes. More complex prompts can be used once you've verified the basic functionality.

Look for the submit button near the text input area — this is typically represented by a send icon (paper airplane), an arrow button, or a button labeled "Send" or "Submit." Alternatively, you can usually press Enter or Ctrl+Enter (Command+Enter on Mac) to submit your prompt without clicking.

Click the submit button to send your test prompt to Grok-4. The interface should immediately show your prompt in the chat history above the input area. You may see a loading indicator such as a spinner, animated dots, or a "thinking" message while Grok-4 processes your request.

The response time typically ranges from a few seconds to about 15-20 seconds depending on server load and prompt complexity. Do not refresh the page or navigate away while waiting for the response.

After a few moments, Grok-4 will generate and display its response in the chat interface. The response should appear below your original prompt in a clearly formatted message box. Read through the response to verify that the model correctly identifies itself as Grok-4 or Grok-4-0709.

A successful response confirms that you've established a working connection to the Grok-4 model through LM Arena. The model may provide additional information about its capabilities, training date, or developer (xAI). This confirmation validates that you can now use Grok-4 for your actual use cases without needing a SuperGrok subscription.

Continue using Grok-4 by typing additional prompts in the same chat interface. Each prompt you submit will be processed by Grok-4, allowing you to use its capabilities for research, creative writing, problem-solving, or any other task. The chat history will remain visible above, allowing you to reference previous exchanges in your conversation.