From Move One to Checkmate, ChatGPT Ruled Them All

In a dramatic conclusion to the AI Chess Exhibition Tournament hosted on Kaggle gaming arena from August 5–7, 2025, OpenAI’s ChatGPT o3 emerged victorious, defeating xAI’s Grok 4 in a dominant 4‑0 sweep to claim the title.

The Final Showdown

The final was a straight contest: ChatGPT dismantled Grok 4 with precision. Grok had initially performed strongly but crumbled under pressure, committing repeated tactical blunders—including significant queen losses. In the third-place playoff, Google’s Gemini 2.5 Pro defeated OpenAI’s o4‑mini with a convincing 3.5‑0.5 score, securing third place. 

More Than Just Gameboard Glory

While the event showcased AI on the chessboard, organizers intended a deeper purpose: testing general-purpose AI models in strategic reasoning—far beyond preprogrammed chess engines.

The Reactions

Magnus Carlsen, watching live, didn’t hold back. He mocked Grok’s performance—likening it to “that one guy who has learnt theory and literally knows nothing else”. According to an estimation done by TechRadar, Grok performed roughly 800 Elo versus o3’s “respectable” 1200 Elo.

Elon Musk, for his part, downplayed the loss, stating that xAI “spent almost no effort on chess,” calling Grok’s win in early rounds a mere “side effect”.

What This Means for AI

This competition underscores a growing shift: general-purpose AI models are now tested not just in language tasks but in structured, rule-based strategy. If models like o3 can outperform others in chess, they may soon take on complex real-world domains like logistics, automated planning, and beyond. 

[1],[2]