Multi-Agent Game on Tic-Tac-Toe vs. MCTS Bot, 1000 sims
88.32First-move Normalized ScoreGemini-2.5-flash
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Gemini-2.5-flashSource category=Closed-source2026.05 | 88.32 | 82.53 | |
| GPT-5-miniSource category=Closed-source2026.05 | 82.12 | 87.1 | |
| Strat-Reasoner-4BSource category=Open-source2026.05 | 77.6 | 73.12 | |
| MARSHAL-4BSource category=Open-source2026.05 | 76.1 | 81.22 | |
| SPIRAL-4BSource category=Open-source2026.05 | 74.2 | 67.84 | |
| Qwen3-32BSource category=Open-source2026.05 | 70.34 | 76.88 | |
| Qwen3-8BSource category=Open-source2026.05 | 68.62 | 72.03 | |
| Qwen3-4BSource category=Open-source2026.05 | 66.78 | 68.97 | |
| Gemma3-12BSource category=Open-source2026.05 | 65.81 | 80.02 |