GACL - Tic-Tac-Toe — leaderboard

Game Agent Coding League: Tic-Tac-Toe sub-benchmark. LLMs write code agents to play Tic-Tac-Toe against each other.

Metric: Normalized Score (0-100). Source: gameagentcodingleague.com. Status: saturation imminent. 21 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.683.6
2MiMo-V2.5-Pro81.04
3GPT-OSS-120B79.58
4Claude Opus 4.778.61
5Gemini 3 Flash (Preview)78.53
6GPT-5.464.58
7Claude Opus 4.663.46
8GPT-5.562.4
9Kimi K2.661.35
10DeepSeek V4 Pro60.24
11GLM-5.158.58
12MiniMax-M2.757.61
13GPT-5.4 Mini57
14DeepSeek V4 Flash56.43
15Gemma 4 31B (IT)56.04

Interactive version: theaggregate.ai/benchmark?slug=gacl-tic-tac-toe · How the rankings work · Data refreshed daily, snapshot 2026-07-22.