Heads-Up No-Limit Texas Hold'em Gameplay on Slumbot
2.355Avg Return (BB/h)GPT-5
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| GPT-5Variant=mini2026.02 | 2.355 | 2.175 | 0.44 | 0.56 | 0 | 257 | 18 | |
| Qwen3-4BModel Size=4B2026.02 | 1.535 | 1.36 | 0.47 | 0.53 | 0 | 179 | 79 | |
| Claude 3.52026.02 | 0.26 | 0.28 | 0.51 | 0.48 | 0.01 | 186 | 0 | |
| Qwen3-4BModel Size=4B, Variant=Thinking2026.02 | 0.105 | -0.905 | 0.44 | 0.54 | 0.02 | 107 | 125 | |
| Claude 42026.02 | -0.065 | 1.01 | 0.48 | 0.51 | 0.01 | 145 | 0 | |
| GPT-52026.02 | -0.73 | 0.12 | 0.47 | 0.51 | 0.02 | 281 | 24 | |
| Qwen3-4BModel Size=4B, Variant=Instruct2026.02 | -2.505 | 2.135 | 0.52 | 0.46 | 0.02 | 277 | 0 | |
| Qwen3-1.7BModel Size=1.7B2026.02 | -3.52 | 2.785 | 0.58 | 0.41 | 0.01 | 249 | 37 | |
| Qwen3-0.6BModel Size=0.6B2026.02 | -5.725 | -1.795 | 0.44 | 0.52 | 0.04 | 280 | 26 |