| Benchmark | Model | Rank | Results |
|---|---|---|---|
| mathematical-reasoning-on-olympiadbench | Qwen3-30B-A3B + ESRL | #22 | Accuracy: 53.5 |
| code-generation-on-livecodebench-v6 | Qwen3-30B-A3B + ESRL | #72 | Pass@1: 0.42 |
| mathematical-reasoning-on-aime24 | Qwen3-30B-A3B + ESRL | #110 | Acc: 24.4 |