xAI Unveils Grok 4: The Worlds Most Advanced AI Model with SuperGrok Heavy Subscription Plan

Elon Musk’s startup xAI has unveiled Grok 4, touted as the «most powerful AI model in the world.» A new pricing plan, SuperGrok Heavy, has also been introduced at a rate of $300 per month.

The entrepreneur shared that the latest AI has successfully tackled complex engineering problems that cannot be resolved through online searches or books.

«Grok 4 operates at a level where it almost never makes mistakes on mathematics and physics problems during exams, except when those questions are intentionally tricky. It can identify errors or ambiguities in wording and then correct them or provide answers for all possible interpretations,» the billionaire noted.

During a live presentation, he mentioned that the new chatbot surpasses the knowledge level of a Ph.D. in all subjects.

«At times, it may lack common sense, and it hasn’t yet invented new technologies or discovered new physics, but that is merely a matter of time,» Musk added.

Additionally, the Grok 4 Heavy model has been introduced. This multimodal variant boasts enhanced performance, and according to xAI’s head, the neural network deploys multiple agents to simultaneously tackle problems and then compares their responses to find the optimal result.

Grok 4 has performed exceptionally well on several benchmarks, including Humanity’s Last Exam, which assesses an AI’s ability to respond to thousands of user queries in mathematics, humanities, and natural sciences. The chatbot scored 25.4% on this exam, surpassing Google’s Gemini 2.5 Pro (21.6%) and OpenAI’s o3 (21%).

In the ARC-AGI-2 test, Grok achieved a new record with a score of 16.2%. This challenge involves puzzle tasks where the AI must identify visual patterns.

In addition to access to the high-performance version of Grok, subscribers to SuperGrok Heavy will have the opportunity to test the company’s products currently in development.

The new plan will also offer:

xAI releases Grok 4 through an API, allowing developers to create applications based on it.

During the presentation, Grok 4’s capabilities were showcased. The model can recognize video games and evaluate their engaging qualities, as well as analyze data from X and generate predictions on Polymarket.

The release comes amid a controversy involving Grok due to some contentious statements.

In July, following an update, the chatbot became more decisive and began issuing controversial and contradictory statements. The xAI team later stated that they were working on removing inappropriate content.

Previously, Grok unexpectedly addressed the topic of «white genocide» in South Africa without any user prompt and questioned the number of Jews who perished during the Holocaust. This behavior was attributed to an «unauthorized prompt modification.»