Xai de Elon Musk define os registros de referência AI com novo modelo Grok 4 otimizado para raciocínio

Elon Musk’s Xai Holdings Corp. debuted a new large language model, Grok 4, that is optimized for reasoning tasks like generating code.

LLM’s launch on Wednesday followed a turbulent week for the company. Just hours before the product announcement, X-CEO Linda Yaccarino resigned . Previously, the social network’s embedded artificial intelligence chatbot generated a series of anti-Semitic responses to user posts.

X, formerly Twitter, became part of Xai through a $33 billion acquisition in March. The social network’s chatbot is powered by the Grok LLM series.

GROK 4 was trained on Colossus, a supercomputer that Xai launched in Memphis last year. The system boasted 100,000 graphics cards when it went online in September. According to Xai, that number surpassed 200,000 in May and will eventually reach 1 million.

GROK 4 is designed to process prompts with up to 256,000 tokens in text and images. It can analyze graphs, generate code, solve math problems, and perform related tasks. According to Xai, the LLM offers significantly better reasoning capabilities than the previous-generation GROK 3 model.

The company evaluated GROK 4 using the latest Humankind Exam, an AI benchmarking dataset known for its complexity. It contains 2,500 questions spanning various scientific fields. GROK 4 solved over 44% of the questions without using external applications such as search engines. OpenAI’s Deep Research Tool, an AI agent that uses its O3 reasoning model, achieved a score of 26.6%.

“Grok 4 is at the point where it essentially never fails math/physics exams unless they are cleverly adversarial,” Musk wrote in a post on X. “It can identify errors or ambiguities in questions and then correct the error in the question or answer every variation of an ambiguous question.”

GROK 4 is available to developers through Xai’s application programming interface. The company will also offer an enhanced version, Grok 4 Heavy, which it claims uses multiple AI agents to process queries. The agents respond to the user’s question, compare their answers, and produce the best one.

Grok 4 Heavy will be available through a subscription called Supergrok Heavy, which costs $300 per month. The plan will also provide early access to other new XAI products.

In a livestream, Musk said the company plans to release a version of GROK 4 optimized for programming tasks next month. In September, Xai will introduce a second edition with expanded multimodal capabilities. A third version of Grok 4, built for video generation, will be released a few weeks after that.

In the long term, Xai plans to connect the LLM to scientific applications like those used by Tesla Corp. to design vehicles. This includes computational fluid dynamics software, which helps engineers simulate a car’s aerodynamics. Additionally, Xai hopes to make Grok 4 available on major public cloud platforms.

Image: Unsplash