| Benchmark | Model | Rank | Results |
|---|---|---|---|
| mathematical-reasoning-on-amc23 | TGRL (Qwen3-14B) (extra data) | #1 | Acc: 92.7 |
| agents-on-alfworld | TGRL (Qwen2.5-7B-Instruct) (extra data) | #4 | Success Rate: 86.7 |
| mathematical-reasoning-on-olympiadbench | TGRL (Qwen3-32B) (extra data) | #8 | Accuracy: 69.9 |
| mathematical-reasoning-on-minerva-math | TGRL (Qwen3-32B) (extra data) | #10 | Accuracy: 50.6 |
| mathematical-reasoning-on-math-500 | TGRL (Qwen3-14B) (extra data) | #24 | Accuracy: 95.3 |
| mathematical-reasoning-on-aime-2025 | TGRL (Qwen3-32B) (extra data) | #85 | Accuracy: 51.6 |