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

DABStep

Benchmarks

Task NameDataset NameSOTA ResultTrend
Data AnalysisDABStep hard
Accuracy37.04
13
Data AnalysisDABStep easy
Accuracy83.33
13
Data AnalysisDABStep 2025 (hard-level)
Accuracy45.24
12
Data AnalysisDABStep 2025 (easy-level)
Accuracy87.5
12
Multi-step Reasoning over Code DependenciesDABstep hard
Accuracy24.34
6
Showing 5 of 5 rows