Seven Grok 4 Examples to Try in the Chat Interface

Learn how to make the most of Grok 4 through seven practical examples you can try in the chat interface.

Jul 17, 2025 · 7 min read

xAI recently released Grok 4, and following the great results of its previous versions, I expected nothing less. In this hands-on blog, I’ll walk you through how to use Grok 4 for practical tasks in the chat interface. We’ll explore how to:

Render an HTML animation
Summarize real-time social media data
Generate an audio simulation
Create a story from an image sketch, and more!

We keep our readers updated on the latest in AI by sending out The Median, our free Friday newsletter that breaks down the week’s key stories. Subscribe and stay sharp in just a few minutes a week:

Example 1: HTML Rendered Animation

I will walk you through a couple of experiments that I tried with Grok 4 in the chat interface to understand the capabilities of this model. If you want to follow along, you’ll need to update to the SuperGrok subscription plan, which gives you access to Grok 4.

This first experiment aimed to recreate an animation showcasing a star's life from its beginning until it reaches the supernova stage. The following prompt is inspired by the animation previewed on the livestream by Grok 4 creators.

Prompt: Build a 30-second HTML animation depicting the stages of a star's life from nebula to supernova, ensuring physical accuracy in the color transitions and expansions.

You can preview the HTML animation directly in the chat interface. The results turned out to be very good and were on par with the simulation I built with Gemini Diffusion.

Example 2: Data Visualization of Complex Networks

Next, I tried visualizing a complex network using the NetworkX library and running the Grok-generated Python script directly in the browser.

Prompt: Use NetworkX in Python to generate and plot a random scale-free network with 100 nodes, highlighting the hubs in red for visual impact.

The above prompt generated the following code:

import networkx as nx
import matplotlib.pyplot as plt

# Generate a random scale-free network using Barabási-Albert model
# n=100 nodes, m=2 (number of edges to attach from a new node to existing nodes)
G = nx.barabasi_albert_graph(100, 2)

# Calculate degrees to identify hubs
degrees = dict(G.degree())

# Sort nodes by degree in descending order
sorted_nodes = sorted(degrees, key=degrees.get, reverse=True)

# Define hubs as the top 10% nodes with the highest degree
num_hubs = int(0.1 * len(G.nodes()))
hubs = set(sorted_nodes[:num_hubs])

# Assign colors: red for hubs, blue for others
node_colors = ['red' if node in hubs else 'blue' for node in G.nodes()]

# Plot the network
plt.figure(figsize=(12, 8))
pos = nx.spring_layout(G, seed=42)  # For reproducible layout
nx.draw(G, pos, node_color=node_colors, with_labels=False, node_size=50, edge_color='gray', alpha=0.7)
plt.title('Scale-Free Network with 100 Nodes (Hubs in Red)')
plt.axis('off')
plt.show()

You can run this code directly in the browser, much like you do with ChatGPT. The code returns an image of a network depicting important hubs in red dots. Thus, Grok4 does a pretty good job at the visualization task as well.

Another interesting capability demonstrated during the livestream was asking Grok to return interesting profile images of X employees. So, inspired by that I tried testing Grok’s web search ability and real-time social media skills.

Prompt: Identify the three most-liked X posts (formerly Twitter) discussing recent major Large Language Model (LLM) announcements. For each identified post, provide the following details: the full text of the post, the date it was published, and the specific LLM announcement it refers to (e.g., model name, new capability, benchmark score). Present this information in a clear and concise format.

It seems like the tech community loves Claude 4, Anthropic’s Project Vend, and India’s Sarvam platform.

Example 4: Live Conversation Simulation

We all work with chatbots, but we want something more conversational, like a one-on-one chat experience or voice conversations. Keeping this thought in my mind, I tried this experiment to simulate a conversation between Grok 4 and a user.

Prompt: Simulate a natural, live conversation where the user casually asks Grok 4 for advice on staying motivated while working from home. Grok 4 should respond with empathy, acknowledging common struggles like isolation or procrastination, and offer a few practical suggestions. As the conversation progresses, the user jokingly asks Grok 4 to write a short motivational rap or jingle. Grok 4 then delivers a fun, rhyming response in a conversational tone. The entire interaction is transcribed live on-screen with subtle animations indicating when Grok 4 is listening, thinking, and responding.

Make sure to have the sound on.

Example 6: Sketch To Story

Inspired by Google Gemini's capabilities of converting a sketch to an image, I tried building a story based on a random sketch and converting the sketch to an image as well.

This feature is available in both Grok 3 and Grok 4. However, the nuance and depth of the story generated by Grok 4 proved to be significantly more compelling than that produced by Grok 3.

Prompt: Build a story around this sketch:

For the visual aspect of this experiment, I then prompted Grok 4 to convert the sketch directly into an image, and I got the following results:

Prompt: Convert this sketch to an image

It’s interesting to note that Grok integrates a watermark into each AI-generated image, enhancing accountability and transparency. However, in terms of raw image generation quality from sketches, I observed minimal difference between Grok 3 and Grok 4.

Example 7: Analyzing Charts

The Grok team emphasized Grok 4's potential for analyzing charts, particularly with its real-time capabilities. To put this to the test, I provided a prompt to analyze a trend chart and supplement its analysis with current information on companies actively investing in AI.

Prompt: Analyze this chart and find more relevant information around companies actively investing in AI

Source: Adaptive AI Market

Grok 4 efficiently processed this task, providing a comprehensive analysis of the chart.

Why Is Grok 4 Important?

Grok 4 evokes the same "wow" factor as Gemini Diffusion with its ability to generate live previews, execute Python scripts, and create dynamic HTML simulations, all from natural language prompts. However, Grok 4 leans more into real-time reasoning, social awareness, and data-driven decision-making.

It’s a highly capable model for processing and responding to real-world inputs, whether that's data scraped from the web or trending posts on X (formerly Twitter). Grok 4 reasons, predicts, and adapts to context in creative and often visually engaging ways.

That said, its current strengths are concentrated in specific domains: scientific simulations, social data search, generative coding, and enterprise dashboard insights. While its context window is limited to 256,000 tokens, significantly smaller than Gemini 2.5 Pro’s 1 million, it still offers a powerful glimpse into what’s possible with the upcoming versions.

If you're a developer aiming to prototype quickly, Grok 4 is highly recommended. However, for production-grade applications, its reasoning capabilities would benefit from being paired with an extended context window in future iterations—learn more in this blog on context engineering.

Conclusion

Grok 4 combines fast, structured reasoning with support for simulations, visual storytelling, audio generation, and interactive coding, all through natural language prompts.

Throughout this tutorial, we explored how Grok 4 can render physics-based animations, visualize complex networks, simulate conversations, and even tackle PhD-level chemistry problems. Its support for tool use, structured output, and real-time web data makes it a promising assistant for developers, researchers, and creative professionals.

To learn more about Grok 4, I recommend these two articles: