Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

TabArena

Benchmarks

Task NameDataset NameSOTA ResultTrend
Binary ClassificationTabArena
Elo Rating1,782
74
Multiclass ClassificationTabArena Lite
Elo Rating1,662
63
RegressionTabArena Lite
Average Rank1.9
56
Tabular ClassificationTabArena Binary
Avg Balanced Accuracy89.32
56
RegressionTabArena Regression
Avg Balanced Acc70.11
56
Multiclass ClassificationTabArena Multiclass (6)
Average Balanced Accuracy0.8109
56
Tabular LearningTabArena
Elo1,663
54
Tabular PredictionTabArena all 51 datasets
Elo Rating1,800
38
RegressionTabarena REG
Average Rank3.46
32
Tabular PredictionTabArena small (full 681 tasks)
Elo Rating1,723
26
Tabular PredictionTabArena medium 10k-100k (full)
Elo Rating2,146
26
RegressionTabArena
Elo Rating1,959
26
Tabular Classification and RegressionTabArena Lite
Mean Accuracy88.7
19
Tabular ClassificationTabArena
Churn Score0.9338
12
Tabular Data LearningTabArena-Lite
Elo Rating1,651
11
Binary ClassificationTabArena Customer v0.1 (test)
ROC AUC0.738
10
Tabular Foundation Model PretrainingTabArena 38 classification tasks subsampled 2025 (5-fold CV)
Time (min)0.92
9
ClassificationTabarena CLS
AUC0.8638
9
Multiclass ClassificationTabArena Multiclass v1 (test)
Rank2.3
6
Binary ClassificationTabArena Binary v1 (test)
Rank2.1
6
ClassificationTabArena All v1 (test)
Rank2.3
6
Tabular ClassificationTabArena (3 outer-fold averages)
Churn Accuracy92.66
4
Showing 22 of 22 rows