Grok 3 is officially released by xAI for Premium+ subscribers after Elon Musk recently announced X. The model has upgraded reasoning, faster response times, and improved creative responses. xAI continues to push its limits, with the very early users testing it in different ways.
10 Wild Examples of Grok 3
Let’s walk through 10 of the wildest uses for Grok 3 :
1) Solving a problem only solved till now by OpenAI’s o1
1. The “Think” button gives it state-of-the-art reasoning to solve problems that many top models struggle with. A user shared his experience of successfully solving complex logic tasks like generating a Settlers of Catan hex grid and generating an SVG of pelican riding a bicycle.
2) Elon Musk shared the wrong answer.
Users have also questioned Elon’s use of the model and how it gives a different answer for a similar prompt when used by different users stating when Elon asked about “The Information”(media company), the model replied it was garbage but when a user put it to test on their own, it conveyed positive emotions about the same.
3) Creating a Game
If you want to create your games now is the time. You can successfully generate a 3D game with the latest version.
For example, a user created one in Pygame inspired by Portal 2 and added a working portal gun. However, it failed to implement sound effects due to missing required audio files.
The prompts given were:
- first prompt: make a 3D game in pygame inspired by portal 2
- second prompt: add a portal gun. right click makes orange portal. left click makes blue portal
- third prompt: add sounds (required sounds files hence failed)
4) Not Good at Coding
However, some users were dissatisfied with the performance and were quick to quip how it wasn’t great at coding as compared to other models, especially when tested on the physics simulation of a ball rolling inside a hexagon compared to o3 mini and Claude 3.5 sonnet.
5) Grok 3 knows Physics
On the other hand, other users thought the physics simulation of a ball rolling inside a hexagon was the best among all models in an even more complex scenario.
They quipped that even without “thinking mode” enabled, it successfully generates Python code for complex physics simulations, like a ball bouncing inside a spinning tesseract. Its potential for real-world physics applications is currently debatable.
6) Grok 3 Made a Video Game
Just like we saw before, users are experimenting by creating different games. Another user, generated an entire playable game with a few prompts in a single file, showing that game development is now faster and more accessible than ever.
7) Tested by Palantir
Despite being newly released, Grok 3 is receiving attention from tech giants like Palantir beating Sonnet and GPT-4o. This shows how the industry is trusting the model and backing it to show strong potential.
8) Elon shared a poem written by Grok 3
We definitely cannot leave out meme culture and creativity. Early users are pushing to analyze old internet posts, generate meme-worthy responses, and even create edits on demand, proving its deep understanding of internet culture (something even Elon finds hilarious).
9) Grok 3 beats flagship Models
Elon has also shared how the model is scoring over 1400 in LM Arena, outperforming leading models like Gemini 2.0 and GPT-4o, making it the highest-rated AI in head-to-head comparisons.
10) Voice Mode Tested
Lastly, for the biggest update, a look at Grok-3’s voice mode shows promising potential, bringing natural and expressive AI-generated speech to the table.
Takeaways
The latest version has gathered varied opinions from users on all forums but one thing we can be sure of is that this is changing the way we use AI and pushing the boundaries for all models. You can try it here!
