Medical Question Answering on HealthBench Medical
56.36ScoreEvoRubric
Evaluation Results
| Method | Links | |
|---|---|---|
| EvoRubricBackbone=Qwen3-14B2026.05 | 56.36 | |
| External Evolving-RLBackbone=Qwen3-14B2026.05 | 54.54 | |
| EvoRubricBackbone=Qwen3-8B2026.05 | 53.97 | |
| Static Rubric-RLBackbone=Qwen3-14B2026.05 | 52.45 | |
| Gemini-2.5-proBackbone=Proprietary2026.05 | 51.07 | |
| External Evolving-RLBackbone=Qwen3-8B2026.05 | 48.77 | |
| Static Rubric-RLBackbone=Qwen3-8B2026.05 | 47.28 | |
| GPT-4oBackbone=Proprietary2026.05 | 46.98 | |
| Base ModelBackbone=Qwen3-14B2026.05 | 46.36 | |
| Base ModelBackbone=Qwen3-8B2026.05 | 42.22 |