GACL - Surround Morris — leaderboard
Game Agent Coding League: Surround Morris sub-benchmark. LLMs write code agents to play a mill/morris board game variant.
Metric: Normalized Score (0-100). Source: gameagentcodingleague.com. Status: saturation imminent. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 79.27 |
| 2 | DeepSeek V4 Pro | 75.47 |
| 3 | Claude Opus 4.7 | 68.44 |
| 4 | GPT-5.4 | 68.07 |
| 5 | Kimi K2.6 | 66.46 |
| 6 | Qwen 3.6 Plus | 63.23 |
| 7 | Gemini 3.1 Pro (Preview) | 60.73 |
| 8 | Claude Opus 4.6 | 53.7 |
| 9 | Claude Sonnet 4.6 | 51.67 |
| 10 | GLM-5.1 | 48.85 |
| 11 | MiniMax-M2.7 | 48.33 |
| 12 | Gemma 4 31B (IT) | 48.18 |
| 13 | DeepSeek V4 Flash | 42.4 |
| 14 | Qwen 3.6 Max Preview | 41.61 |
| 15 | MiMo-V2.5-Pro | 37.86 |
Interactive version: theaggregate.ai/benchmark?slug=gacl-surround-morris · How the rankings work · Data refreshed daily, snapshot 2026-07-22.