Tabular Classification on Benchmark runs Macro-mean
90.8AUCTabICLv2
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TabICLv2Hardware=GPU (A100-class), Model Role=TFM teacher2026.05 | 90.8 | 151 | |
| TabPFNv2.6Hardware=GPU (A100-class), Model Role=TFM teacher2026.05 | 90.2 | 327 | |
| LimiXHardware=GPU (A100-class), Model Role=TFM teacher2026.05 | 90 | 448 | |
| TabICLv2→XGBHardware=single CPU core, Model Role=Distilled student2026.05 | 88.2 | 1.9 | |
| TabICLv2→CBHardware=single CPU core, Model Role=Distilled student2026.05 | 88.2 | 2.7 | |
| OrionMSP v1.5Hardware=GPU (A100-class), Model Role=TFM teacher2026.05 | 87.8 | 1,275 | |
| TabICLv2→LGBMHardware=single CPU core, Model Role=Distilled student2026.05 | 87.8 | 4 | |
| CatBoostHardware=single CPU core, Model Role=Baseline2026.05 | 87.6 | 1.2 | |
| LightGBMHardware=single CPU core, Model Role=Baseline2026.05 | 87.2 | 1.4 | |
| TabICLv2→MLPHardware=single CPU core, Model Role=Distilled student2026.05 | 84.2 | 1.5 |