The first chess tournament between advanced artificial intelligence models saw OpenAI’s o3 model win a decisive victory over Grok 4 from xAI, a company founded by Elon Musk.
In the final series of the tournament on the Kaggle AI Exhibition Tournament platform, OpenAI o3 won 4-0. Both models made it to the final, beating competitors from Anthropic, DeepSeek, Google and Moonshot AI. This match is particularly symbolic, given that the heads of OpenAI and xAI, Sam Altman and Elon Musk, co-founded OpenAI ten years ago, but Musk then left to create his own AI project.
The significance of the chess match for artificial intelligence
Chess has long been a benchmark for AI development. Back in 1997, IBM’s Deep Blue supercomputer defeated then-champion Garry Kasparov, and later projects such as Google’s DeepMind created machines capable of learning independently and achieving a level exceeding that of humans.
This new tournament is a unique test for widely used large language models (LLMs), which are currently undergoing rapid development following the launch of ChatGPT in 2022. According to chess champion Magnus Carlsen, Grok 4 has a rating of approximately 800 points, and o3 has a rating of approximately 1,200, which is much lower than the level of the strongest grandmasters, but interesting for understanding the capabilities of modern AI systems.
Elon Musk noted that Grok’s chess abilities are more of a side effect, as the xAI team focused on other tasks. Nevertheless, this match demonstrates the growing capabilities of artificial intelligence in various fields.
“This is an important step towards better understanding how modern AI models work and reason,” an AI expert said.
This tournament opens up new horizons for research and application of artificial intelligence, showing that LLM can not only understand language, but also solve complex tasks that require strategic thinking.
