Multi-hop Question Answering on LB QA v1
77.3AccuracyQwen3-30B-A3B-Thinking-2507 + Data Recipe for Long-Context RL
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-30B-A3B-Thinking-2507 + Data Recipe for Long-Context RLBase Model=Qwen3-30B-A3B-Thinking-2507, Training Recipe=Data Recipe for Long-Context RL2026.06 | 77.3 | |
| Qwen3-30B-A3B-Thinking-2507 + DocQA-RL-1.6KBase Model=Qwen3-30B-A3B-Thinking-2507, Training Recipe=DocQA-RL-1.6K2026.06 | 68.9 | |
| Qwen3-8B-128K + Data Recipe for Long-Context RLBase Model=Qwen3-8B-128K, Training Recipe=Data Recipe for Long-Context RL2026.06 | 68.8 | |
| Qwen3-30B-A3B-Thinking-2507 + KeyChain-15KBase Model=Qwen3-30B-A3B-Thinking-2507, Training Recipe=KeyChain-15K2026.06 | 68.4 | |
| Qwen3-4B-Thinking-2507 + Data Recipe for Long-Context RLBase Model=Qwen3-4B-Thinking-2507, Training Recipe=Data Recipe for Long-Context RL2026.06 | 68.3 | |
| Qwen3-4B-Thinking-2507 + KeyChain-15KBase Model=Qwen3-4B-Thinking-2507, Training Recipe=KeyChain-15K2026.06 | 66.8 | |
| Qwen3-8B-128K + KeyChain-15KBase Model=Qwen3-8B-128K, Training Recipe=KeyChain-15K2026.06 | 66.5 | |
| Qwen3-30B-A3B-Thinking-2507Base Model=Qwen3-30B-A3B-Thinking-2507, Training Recipe=Base2026.06 | 66.5 | |
| Qwen3-4B-Thinking-2507 + DocQA-RL-1.6KBase Model=Qwen3-4B-Thinking-2507, Training Recipe=DocQA-RL-1.6K2026.06 | 63.8 | |
| Qwen3-8B-128K + DocQA-RL-1.6KBase Model=Qwen3-8B-128K, Training Recipe=DocQA-RL-1.6K2026.06 | 63.7 | |
| Qwen3-8B-128KBase Model=Qwen3-8B-128K, Training Recipe=Base2026.06 | 62.9 | |
| Qwen3-4B-Thinking-2507Base Model=Qwen3-4B-Thinking-2507, Training Recipe=Base2026.06 | 62.6 |