Artificial Intelligence

OpenAI's o3 Defeats Musk's Grok in AI Chess Tournament

Published AUG 8 2025

Story

Image credit: Future Publishing/Getty Images

The Facts

OpenAI's o3 model defeated xAI's Grok 4 with a 4-0 victory in the final of Kaggle's AI Chess Exhibition Tournament held on Thursday, marking the first major tournament testing general-purpose large language models (LLMs) in chess competition.
Chess.com
The Independent
The tournament featured eight LLMs from companies including Anthropic, Google, OpenAI, xAI, DeepSeek and Moonshot AI, who competed over three days on Google-owned platform Kaggle to assess their reasoning and strategic capabilities.
Mobile World Live
BBC News
Google's Gemini 2.5 Pro secured third place by defeating OpenAI's o4-mini model by a score of 3.5-0.5 in the bronze medal match, while o3 remained unbeaten throughout the entire tournament, according to competition organizers.
Chess.com
Mobile World Live
Grok 4 made multiple errors during the final, including repeatedly losing its queen and dropping pieces for no apparent reason, with chess grandmaster Hikaru Nakamura noting the AI made numerous mistakes during his livestream commentary.
Chess.com
BBC News
The competition used chess as a benchmark to measure AI models' abilities in strategic reasoning, long-term planning and dynamic adaptation, with Kaggle explaining games offer a clear signal of success for evaluating AI performance.
Tech Digest
Mobile World Live
Elon Musk posted on X that xAI spent almost no effort on chess and described Grok's chess ability as a "side effect," while the tournament adds fuel to his ongoing rivalry with OpenAI co-founder Sam Altman, who recently dismissed Musk's criticisms, claiming he thought he was just "tweeting all day."
BBC News
Cryptopolitan
Elon Musk

Sources Split

LEFT

CENTER

RIGHT

ESTAB
CRIT

CENTER

PRO
ESTAB

The Spin

Narrative A

This chess beat down proves Grok is officially no match for OpenAI. Even world chess champ Magnus Carlsen mocked Grok's bad moves. Besides that, ChatGPT is faster, smarter and better at math, coding and handling text, images and voice. With 400 million weekly users, ChatGPT rules, while Grok's social media tricks can't keep up. This loss shows Grok's sliding behind.

Nitro Media Group

Firstpost

Narrative B

Grok 4's 4-0 chess loss to ChatGPT's o3 doesn't spell doom. It still shines in complex reasoning, math and coding, with a multi-agent system that tackles tough problems like a think tank. At $30/month, it's also cheaper than ChatGPT-5's $200 Pro tier, and its edgy, rebellious tone sparks creative ideas. With real-time X data, Grok 4 remains a strong contender for innovative businesses, far beyond a mere chess match.

FOX News

McNeece Web Design

Techno-skeptic narrative

While the rapid rise in machine learning has been fun, from Deep Blue defeating Garry Kasparov in chess in 1997 to the launch of chatbots, AI has now become dangerous. The likelihood of an AI-induced catastrophe has grown exponentially, either from malicious human actors or a sentient AI itself, and it's time for the world to take this threat seriously and demand strict safeguards before it's too late.