GACL - Connect4: leaderboard
Game Agent Coding League: Connect4 sub-benchmark. LLMs write code agents to play Connect Four against each other.
Metric: Normalized tournament points (0-100); source-selected best generated agent. Source: gameagentcodingleague.com. Saturation forecast: Around January 2027. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 76.3 | #58 |
| 2 | Claude Opus 4.7 | 73.23 | #49 |
| 3 | GLM-5.1 | 71.2 | #97 |
| 4 | GPT-5.5 | 69.64 | #31 |
| 5 | GPT-5.4 | 69.32 | #79 |
| 6 | Kimi K2.6 | 66.46 | #104 |
| 7 | Claude Opus 4.6 | 65.57 | #63 |
| 8 | Claude Sonnet 4.6 | 57.81 | #90 |
| 9 | Qwen 3.6 Max Preview | 56.41 | #85 |
| 10 | Qwen 3.6 35B A3B | 55.83 | #226 |
| 11 | Qwen 3.6 Plus | 55 | #120 |
| 12 | DeepSeek V4 Pro | 54.38 | #102 |
| 13 | MiMo-V2.5-Pro | 45.05 | #158 |
| 14 | Qwen 3.6 27B | 43.54 | #170 |
| 15 | MiniMax-M2.7 | 42.76 | #258 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=gacl-connect4 · How It Works · Data refreshed daily, snapshot 2026-10-20.