Science Question Answering on ResearchQA Science (Score)
77.31ScoreEvoRubric
Evaluation Results
| Method | Links | |
|---|---|---|
| EvoRubricBackbone=Qwen3-14B2026.05 | 77.31 | |
| EvoRubricBackbone=Qwen3-8B2026.05 | 76.98 | |
| Static Rubric-RLBackbone=Qwen3-14B2026.05 | 76.97 | |
| External Evolving-RLBackbone=Qwen3-14B2026.05 | 76.8 | |
| Static Rubric-RLBackbone=Qwen3-8B2026.05 | 74.93 | |
| External Evolving-RLBackbone=Qwen3-8B2026.05 | 74.7 | |
| Base ModelBackbone=Qwen3-14B2026.05 | 68.85 | |
| Gemini-2.5-proBackbone=Proprietary2026.05 | 68.84 | |
| Base ModelBackbone=Qwen3-8B2026.05 | 67.55 | |
| GPT-4oBackbone=Proprietary2026.05 | 65.98 |