The OpenAI 03 artificial intelligence model from ChatGPT’s developer won the artificial intelligence chess tournament without a single defeat. This was reported by the BBC.
In the final, OpenAI 03 outperformed Grok 4 from Elon Musk’s xAI. Google’s Gemini model took third place, defeating another OpenAI model.
The artificial intelligence chess tournament was held on the Kaggle platform, which is owned by Google and allows data processing specialists to evaluate their systems through competitions. Eight major American language models from Anthropic, Google, OpenAI, xAI, as well as China’s DeepSeek and Moonshot AI, competed against each other during the three-day Kaggle tournament.
Historically, technology companies have often used chess to evaluate the progress and capabilities of computers. Modern chess machines at maximum power are unbeatable even against the best human players.
But computers designed for playing chess did not participate in these competitions. Instead, they were held between artificial intelligence programs designed for everyday use. These artificial intelligences, although leading and talented in many everyday tasks, are still improving at chess. For example, Grok made several mistakes during the final games, including repeatedly losing its queen.
Musk and Sam Altman, both co-founders of OpenAI, claim that their latest models are the smartest in the world.
“Grok made many mistakes in these games, while OpenAI did not,” said Grandmaster Hikaru Nakamura during his live broadcast of the final.
Before the final, Musk wrote on X that xAI’s previous success in the tournament was a “side effect” and that the company had “put almost no effort into chess.”