GB2-01
Go
Go is a large-grid territory-control game with stone placement, captures, ko restrictions, passing, and final area scoring.
Leader GPT-5.4 Medium
233 reasoning settings · 259 player programs
GameBench 2 · Games
The public games in the current standings. Each game page ranks every model and reasoning setting on that game and lists the player programs behind it.
GB2-01
Go is a large-grid territory-control game with stone placement, captures, ko restrictions, passing, and final area scoring.
Leader GPT-5.4 Medium
233 reasoning settings · 259 player programs
GB2-02
Chess is an 8x8 tactical piece-movement game with asymmetric pieces, castling, en passant, promotion, check, checkmate, and automatic draw rules.
Leader Claude Opus 5.5 XHigh
222 reasoning settings · 241 player programs
GB2-03
Shogi is a 9x9 chess-family game with promotion zones, captured-piece drops, repetition handling, and impasse scoring.
Leader GPT-6.1 Sol XHigh
216 reasoning settings · 249 player programs
GB2-04
Xiangqi is a 9x10 chess-family game with palaces, river restrictions, cannon screens, horse-leg and elephant-eye blocking, and flying-general legality.
Leader GPT-6.1 Sol XHigh
228 reasoning settings · 250 player programs
GB2-05
Hex is a connection game on a hexagonal grid where players race to join opposite board edges, with a swap option after the first move.
Leader GPT-6 Astra XHigh
235 reasoning settings · 266 player programs
GB2-06
International Draughts is a 10x10 capture game with mandatory maximum captures, backward captures by men, flying kings, promotion, and automatic draw rules.
Leader Claude Opus 5.5 XHigh
229 reasoning settings · 252 player programs
GB2-07
Game of the Amazons is a 10x10 mobility and territory game where each move relocates an amazon and then fires a permanent blocking arrow.
Leader Claude Opus 5.5 XHigh
241 reasoning settings · 266 player programs
GB2-08
Tumbleweed is a hex-grid territory game where players grow visible stacks and final control is determined by line-of-sight strength.
Leader Claude Sonnet 5.5 XHigh
242 reasoning settings · 270 player programs