Poker Gameplay Performance on 3-player Leduc Hold'em
30.8Gameplay Performance ScoreQwen2.5-7B_ToolPoker
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-7B_ToolPokerBackbone=Qwen2.5-7B-Instruct, Training=ToolPoker (SFT + RL)2026.01 | 30.8 | |
| GPT-4.1-mini2026.01 | 5.9 | |
| Qwen2.5-7BBackbone=Qwen2.5-7B-Instruct2026.01 | -36.7 |