Is Musk’s Grok 3 really the AI game-changer he claims?

Multi-billionaire Elon Musk said on Thursday his AI chatbot, and ChatGPT challenger, Grok 3, is in the final stages of development and will be released within a week or two.

"Grok 3 has very powerful reasoning capabilities, so in the tests that we've done thus far, Grok 3 is outperforming anything that's been released, that we're aware of, so that's a good sign," he said during a video call at the World Governments Summit in Dubai.

As one of the co-founders of OpenAI, Musk said his company had offered $97.4 billion to buy OpenAI's nonprofit assets, in another salvo from the world's richest man against the artificial intelligence startup on Monday.

In late December, OpenAI said it wants to become a for-profit organisation to secure the capital needed for developing the best AI models.

Soon after, Musk filed a lawsuit against OpenAI CEO Sam Altman and others in August, requesting a US district judge to prevent OpenAI from shifting to a for-profit model. OpenAI responded this week, stating that Musk’s legal action contradicts his own lawsuit.

“We shouldn't be surprised that tech companies adopt practices that are designed to extract greater profits,” says Dr Natasha Tusikov, an Assistant Professor of Criminology at York University, commenting on OpenAI’s for-profit model to TRT World.

Musk also described Grok 3 as "scary smart" with "very powerful reasoning capabilities," and emphasised its reliance on synthetic data to achieve logical consistency.

In a January 4 post on X, he also highlighted that Grok 3's pre-training was complete with 10 times more computational power than Grok 2.

Despite Musk’s bold claims, Grok 3’s actual performance remains uncertain.

Benjamin De Kraker, an xAI engineer who worked on the human data team for Grok development, ranked Grok 3 below OpenAI’s models for coding ability in an X post.

He claimed that OpenAI’s o1-pro, o1, and o3-mini are all tied for the top spot, with Grok 3 in fourth position.

Following his post, xAI reportedly issued De Kraker an ultimatum to delete the post or face being fired. Instead he chose to resign, expressing disappointment at xAI's attempt to suppress his opinion.

Musk later called the situation "weird," but he did not indicate whether he would intervene.

Will Grok 3 differ from its rivals?

Dr. Alan D. Thompson, an AI expert specialising in the augmentation of human intelligence emphasises Grok’s unique advantage—its ability to pull real-time data from platforms like X.

According to him, this feature sets it apart from competitors, offering fresh insights and potentially enhancing user experience with continuously updated information.

Thompson also explains that Grok-3 utilises several specialised data collections to enhance its capabilities.

These include DeepMind’s MassiveText for understanding and generating content in multiple languages, EleutherAI’s The Pile for a diverse range of topics, and Hugging Face’s FineWeb, which helps Grok-3 learn from well-written and high-quality information.

However, real-time data access also poses challenges, particularly regarding misinformation. With Grok's less restrictive content policies, the risk of spreading unverified information remains a key ethical concern.

“Grok-3 is one of the main frontier models alongside GPT, Claude, Gemini, and Llama,” says Thompson.

While Musk’s assertions about Grok 3’s superiority sound ambitious, they are based on internal testing rather than independent evaluations.

As the curtain slowly lifts on Grok-3's full potential, the world eagerly awaits to see if it can truly live up to the hype. Its capabilities will only become clear once more people can use it.