Chessnut Move Leela Chess Zero Robot vs GPT 5.5 Thinking Extended 🟡 Gadgetify
LLMs are getting a lot better in chess. They used to make illegal moves all the time but the latest models now can play well with fewer mistakes. According to the latest ChessBench, Gemma 4 can now play at 2537. It is only behind GPT 5.5 Pro, which is rated at 2630 ELO. We recently had the chance to test Stockfish 18 against GPT 5.5 Thinking extended, and while it did lose, it managed to play a respectable game.
The latest ChessBench results are out. ✨
It’s fun to see Gemma 4 ranking in the top 3! pic.twitter.com/nLvTnUnDZJ
— Google Gemma (@googlegemma) May 12, 2026
