Elon Musk's Grok Balayé 4–0 by Openai at chess

On the chessboard, two visions of the AI clashed. Sam Altman, boss of Openai, and Elon Musk, founder of Xai, crossed their models in a chess tournament organized by Google. For three days, the O3 of Openai and Grok 4 from XAI measured themselves without any specialized assistance. Much more than a simple exhibition match, the event turned to the revealer: behind the final score, it is the real gap between two artificial intelligences, and two strategies, which appeared in the open.

A one -way finish

During the final disputed on August 7, the O3 model of Openai, however depreciated a few days earlier for the benefit of GPT-5 which already disappointed, inflicted a 4–0 without appeal to XAI Grok 4.

This tournament, baptized “Kaggle Game Arena Ai Chess Exhibition”prohibited any use of a chess engine or a dedicated training, leaving the models to manage with their general knowledge gleaned on the internet.

From the first parts, the difference was felt. Magnus Carlsen, world champion and commentator of the event, compared the two AI to “A gifted child who does not know how the parts move”believing their level around 800 Elo, far from competitive standards.

Grok clumsiness has multiplied throughout the final:

• Free loss of important parts, offering an immediate material advantage in Openai;
• A failure of a “Poisoned pawn”with a bad target chosen which led to the immediate capture of his lady;
• Solid positions spoiled in the middle of the game, with a succession of incoherent blows;
• Maltus management of the initial advantage in the fourth part, allowing O3 to reverse the situation.

Hikaru Nakamura, Grand Master International and Streamer of the event, summed up the difference between the two opponents: “Openai did not make the mistakes that Grok made”. He also praised O3's spectacular ascent in the last part, where a bad start had nevertheless presented a possible victory for Xai d'Elon Musk.

The limits of generalist AIs exposed

Beyond the score, the tournament revealed the structural difficulties of generalist AIs when they are faced with a strict framework like that of failures. Many models have been disqualified in the preliminary phase after trying impossible actions: teleportation of parts, resurrections of captured units, illegal trips of pawns.

Even in the final, understanding the rules seemed fragile, alternating between brilliant blows and absurd decisions. As Carlsen pointed out, “These AI know how to count the captured parts, but not how to conclude a winning part”.

This observation is not isolated. Earlier this year, the international master Levy Rozman had organized a similar tournament where language models had multiplied illegalities, until invoking disappeared pieces. Stockfish, a specialized AI, had won the high -hand event. These episodes show that, despite the promises and speeches on their versatility, language models remain far from mastering tasks requiring coherence and procedural rigor.

For Elon Musk, this defeat against Sam Altman, the second in direct competition this year, falls at the wrong time, when Xai has just raised $ 10 billion and seeks to position himself as a credible actor in the general AI race. However for the entire sector, Google’s exhibition recalls above all that the major current models excel in the treatment of natural language, much less in the strict application of complex rules. The AI may one day be able to compete with the best chess players … But that day, it will also have proven that it is capable of reasoning far beyond the black and white boxes.