Earth Science Reasoning on EarthSE (test)
58Initial ScoreQWEN3-MAX
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| QWEN3-MAXModel Scale=Max2026.05 | 58 | 0.8 | 2 | 1.2 | 58.8 | |
| SFTModel Scale=9B2026.05 | 48.8 | 0 | 1.6 | 1.6 | 48.8 | |
| CRITIQUEGRPOModel Scale=9B2026.05 | 48 | 2 | 6 | 4 | 50 | |
| GRPOModel Scale=9B2026.05 | 46.8 | 1.2 | 1.6 | 0.4 | 48 | |
| ReCritModel Scale=9B2026.05 | 46.4 | 9.6 | 13.6 | 4 | 56 | |
| DPOModel Scale=9B2026.05 | 45.6 | 0.8 | 2.8 | 2 | 46.4 | |
| BASE MODELModel Scale=9B2026.05 | 44.4 | 0.4 | 3.6 | 3.2 | 44.8 | |
| DPOModel Scale=4B2026.05 | 41.2 | -1.2 | 0.4 | 1.6 | 40 | |
| CRITIQUEGRPOModel Scale=4B2026.05 | 40.8 | 3.6 | 5.2 | 1.6 | 44.4 | |
| SFTModel Scale=4B2026.05 | 40.4 | -0.8 | 0.8 | 1.6 | 39.6 | |
| GRPOModel Scale=4B2026.05 | 40 | -0.8 | 1.2 | 2 | 39.2 | |
| BASE MODELModel Scale=4B2026.05 | 39.2 | -1.6 | 0 | 1.6 | 37.6 | |
| ReCritModel Scale=4B2026.05 | 38.8 | 19.2 | 22.8 | 3.6 | 58 |