Science Question Answering on GPQA Diamond (Pass@1, #Tokens)
57.32Pass@1Qwen3-8B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-8BType=BASE LLM2025.10 | 57.32 | 8,624 | |
| BeConciseType=SHORT COT2025.10 | 57.23 | 7,466 | |
| MixReasoningType=ADAPTIVE2025.10 | 57.17 | 6,734 | |
| O1-PrunerType=SHORT COT2025.10 | 56.95 | 8,543 | |
| NoWaitType=SHORT COT2025.10 | 56.67 | 6,575 | |
| DEERType=SHORT COT2025.10 | 55.45 | 7,778 | |
| CoT-ValveType=SHORT COT2025.10 | 40.76 | 5,317 | |
| ThinklessType=HYBRID2025.10 | 24.75 | 8,013 |