Late on Wednesday, Elon Musk’s artificial intelligence startup, xAI, unveiled SuperGrok Heavy, a new $300/month AI subscription plan, along with its newest flagship AI model, Grok 4.
The surprising part, or maybe not, but many people are claiming that it is smarter than any other AI model.
Is it really? Let’s see what it is and how it is smarter than any AI model.
What Is Grok And Grok4?
Grok is a generative AI chatbot made by xAI, which is a company started by Elon Musk. It launched in November 2023 and is built on a large language model which is also called Grok.
Grok is connected to X and has apps for iOS and Android. The name “Grok” comes from a word in the 1961 science fiction book Stranger in a Strange Land by Robert A. Heinlein. The word means to understand something deeply.
The bot is known for giving “unfiltered answers,” which have sometimes led to controversial replies. It has three versions:
- Grok 1
- Grok 1.5
- Grok 1.5V
- Grok 4
- Grok 4 Heavy
Grok 4 is the latest and smartest version yet. It can use built-in tools and search the internet in real time. Grok 4 is available to SuperGrok and Premium+ subscribers, and also through the xAI API.
Why Grok 4 is smarter And Better?
“With respect to academic questions, Grok 4 is better than PhD level in every subject, no exceptions,” said Elon Musk during a livestream Wednesday night.
Grok 4 thinks more like a human. It can solve problems it has never seen before, and the answers are not just copied from the internet.
Elon Musk also claimed that Grok 4 is smarter than nearly all graduate students in almost every subject at the same time.
Grok 4 is the first time, in my experience, that an AI has been able to solve difficult, real-world engineering questions where the answers cannot be found anywhere on the Internet or in books.
And it will get much better.
Smarter Problem Solving with Real-Time Tools
With Grok 3, the model is trained on predicting the next word at a much bigger scale than ever before. This gave it better knowledge and performance. In fact, Grok 3 was also built on Reasoning, which used reinforcement learning to help the model think longer and solve problems more accurately.
While on GROK 4, the use of reinforcement learning training is much bigger.
For Grok 4, there was a use of Colossus, which is a huge system with 200,000 GPUs, to train the model even more. This helped Grok improve its reasoning skills at a massive scale. This was made possible by:
- Building new infrastructure
- Creating better algorithms that made training 6 times more efficient
- Collecting a lot more high-quality data, not just in math and coding but in many other areas
Thanks to all this, Grok 4 kept getting better during training, using much more computing power than ever before.
Advanced Search and Media Understanding Inside X
Grok 4 was also trained to use tools on its own with reinforcement learning. This helps it solve problems that are normally hard for AI models.
For example, Grok 4 can:
- Use a code interpreter to solve coding tasks
- Search the web in real time to find updated or detailed info
- Pick its own search terms and dig deep online to give high-quality answers
It can also search within X, using smart keywords and meaning-based search tools. Grok can even look at pictures and media to improve its responses. This makes it much better at answering tough or complex questions.
What Can You Expect More In The Future?
In the future, the company is committed to making reinforcement learning even better.
At the moment the task are being simple and controlled, but they want to take it to the nxt levl by solving real-world problems. At this stage, the AI can learn and adapt as things change. Mindblowing!
They will also be improving multimodal abilities, so in the future, the Grok will work with:
- Text
- Images
- Sounds
Overall, the company’s goal is to create models that are:
- Smarter
- Faster
- Efficient
The company is trying to help people in meaningful ways.
What do critics Have To Say?
Grok 4 has the ability to beat AI systems in many tests, for example like in STEM subjects like math and science.
A critic tried it out himself and said:
“Grok did really well with math and programming in my tests. I was impressed by how clearly it explained its thinking and solved problems in a smart, logical way.”
However, he also pointed out some weaknesses:
The context window isn’t very big, so Grok has trouble working with very large amounts of information, like big code projects.
It struggled to analyze a 170-page PDF, probably because of its smaller context window and weaker multimodal skills.
Conclusion
Grok 4 is faster, smarter, and better at solving problems than most machines available right now. No doubt it has gone through massive training, advanced tools, and real-time search abilities. But there are still some limitations to it.
But AI will keep on evolving, and Grok 4 is also going to have more capability, just like handling text, images, and audio.
Still curious? What is going to happen in the future? You are at the right spot! Subscribe to our newsletter for more such interesting updates.
FAQs
How much is Grok 4?
Grok 4 is released by Elon Musk’s xAI with a $300 monthly subscription.
Is Grok 4 free?
Grok 4 is free for testing.
A few days ago, Elon Musk released GROK 4, the best AI in the world, and the internet is exploding with excitement. Unfortunately, not everyone can use the model because it is paid for, falling under a $30/month subscription. This video cannot play in your browser. There was a mistake.
Is Grok better than GPT 4?
Although some users have noted that the comparisons may involve cherry-picking baseline scores from rival models, Grok 4 is outperforming Claude Opus, Gemini 2.5 Pro, and GPT-4 (o3) in the majority of categories.
