OpenAI, the company behind ChatGPT, defeated Grok, owned by Elon Musk, in the a chess championship match of the greatest artificial intelligence (AI) chess player.
Since current chess machines are nearly invincible against even the best human players, tech corporations have historically and frequently used chess to evaluate a computer’s progress and abilities.
But this competition was not between computers made to play chess instead and it was fought between AI systems meant for everyday use.
An already-existing rivalry between the two companies was heightened when OpenAI’s o3 model won the event and defeated xAI’s Grok 4 model in the championship match which was intended for.
The Co-founders of OpenAI Musk and Sam Altman assert that their most recent models are the most intelligent and smart in the world.
After defeating another OpenAI model, Google’s model Gemini took third place in the same tournament.
Even though these AIs are skilled at many everyday jobs, they are still becoming and improving better at chess; in its last games, Grok made several mistakes, including losing its queen several times.
Pedro Pinhata, a journalist for Chess.com, stated in his coverage in which he was opportune to that “it seemed like nothing would be able to stop Grok 4 on its way to winning the event” until the semi-finals.
“X’s AI appeared to be by far the best chess player, despite a few weak spots… However, on the last day of the competition, the illusion was destroyed.
The journalist had claimed that a string of “convincing wins” were made possible by Grok’s “unrecognisable” and “blundering” performance.
During his broadcast of the final, chess champion and grandmaster Hikaru Nakamura pointed out that OpenAI did not make as many mistakes in these games as Grok did.
Musk said in a post on X before to Thursday’s final that xAI “spent almost no effort on chess” and that its preceding tournament success had been a “side effect.”
What makes AI a chess player is the fact that the Google-owned platform Kaggle, which enables data scientists to test their systems through tournaments, hosted the AI chess game.
During Kaggle’s three-day tournament, eight sizable language models from Anthropic, Google, OpenAI, xAI, and Chinese developers DeepSeek and Moonshot AI competed against one another.
Tests known to be benchmarks were used by AI developers to assess the thinking and coding abilities of their models.
The competition which had lasted three days and pitted eight large language models (LLMs) from Chinese developers DeepSeek and Moonshot AI against those from Anthropic, Google, OpenAI, and xAI.
Chess and Go are two instances of a tricky rule-based strategy games that are frequently used to evaluate a model’s capacity to learn how to best accomplish a particular goal, in this case, outwitting opponents to win.
In the latter part of the 2010s, AlphaGo, a computer program created by Google’s AI division DeepMind to play the Chinese two-player strategic game Go, claimed a string of wins over human Go champions.
Lee Se-dol, a South Korean go master, retired in 2019 following multiple AlphaGo defeats.
Telling the Yonhap news agency, “There is an entity that cannot be defeated,” he said.
Sir Demis Hassabis, one of the co-founders of DeepMind, was once a chess prodigy and an expect.
Meanwhile, chess champions faced off against strong computers in the late 1990s. An example of this was the IBM’s Deep Blue defeating Garry Kasparov, the world champion at the time, in 1997, computers have been able to defeat the best human players over time..
The triumph of Deep Blue was regarded as a significant event that showed how well computers could mimic some human abilities.
Twenty years later, Mr. Kasparov compared its intelligence to that of an alarm clock, although he added that he didn’t feel any better after losing to a $10 million (£7.6 million) alarm clock.
Magnus Carlsen, a top player in the world, defeated ChatGPT in an online chess encounter in July without dropping a single piece.
During the final, Carlsen also voluntarily offered live commentary, comparing the games to “watching children’s games.” You play them out in such competitions every time.
He hope everyone feels better about their games after watching this, he joked after the final.
It has been alleged that Carlsen placed Grok’s playing strength to be at about 800 Elo and OpenAI’s o3 at about 1,200, both of which were significantly lower than his own peak rating of 2,882.
We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it.
