Human Preference Alignment on Multi-Challenge
49.4Avg@3Qwen3-30A3-2507
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-30A3-2507Release Identifier=25072025.12 | 49.4 | |
| Qwen3-4B-2507Release Identifier=25072025.12 | 41.8 | |
| Nanbeige4-3B-ThinkingRelease Identifier=25112025.12 | 41.2 | |
| Qwen3-32B-2504Release Identifier=25042025.12 | 39.2 | |
| Qwen3-14B-2504Release Identifier=25042025.12 | 36.4 | |
| Qwen3-8B-2504Release Identifier=25042025.12 | 35.8 |