AI models chess rating list published

Kaggle Game Arena organized the first AI exhibition chess tournament. A unique and powerful approach to understanding AI capabilities, Kaggle Game Arena, in partnership with Google DeepMind, announced a new platform where leading AI models will compete head-to-head in chess and other games. It will feature Kimi k2 (Moonshot AI), Gemini 2.5 Pro (Google), Gemini 2.5 Flash (Google), o3 (OpenAI), o4-mini (OpenAI), Claude 4 Opus (Anthropic), Grok 4 (xAI), and DeepSeek R1.

o3 OpenAI won the Kaggle Game Arena AI Exhibition Chess, but there is more to the story. The Kaggle Game Arena organizers say, “In the first #KaggleGameArena — Chess Text Input — AI models faced off using only text inputs (no tools, no move validation) in 40+ matches per pairing to build a robust Elo-like ranking. We also hosted a 3-day chess exhibition tournament featuring a few streamed matchups, but most games were run behind the scenes to build the final leaderboard.” The estimated human rating of the winner o3 Open AI is 1686 elo. The runner-up Grok is rated 1395 elo. Scroll down for the full table.

Kaggle Game Arena AI chess timeline

03.08.2025 TCEC Season 28 Premier Division started
04.08.2025 Kaggle Game Arena is starting with an AI exhibition chess tournament
04.08.2025 Kaggle AI Chess Exhibition Tournament LIVE
05.08.2025 Gotham Chess, “Google notice me and created the Kaggle Game Arena”
05.08.2025 Interview with Grok ahead of Kaggle Game Arena AI exhibition chess tournament
05.08.2025 Kaggle AI Chess Exhibition pairings
07.08.2025 o3 OpenAI wins Kaggle Game Arena AI Exhibition Chess

Chess rating of AI models

To enter the Top Chess Engine Championship, the premier event for computer chess, a minimum of 3000 elo estimated is required. Yet, the leap of chess strength for LLMs is significant in the past years and it will be interesting to see their development in this field in the near future.