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

SciBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
ChemistrySciBench
Accuracy66.4
32
Physics problem solvingScibench fund
Accuracy74.6
24
Scientific Problem SolvingSciBench-107
Atkins Score62.5
24
Uncertainty CalibrationSciBench
AUROC78.7
18
Scientific Problem SolvingSciBench
Pass@2077.46
17
Scientific ReasoningSciBench
Score28.52
17
Scientific ReasoningSciBench
Pass@127.9
10
Scientific ReasoningSciBench
Pass@1072.23
6
ScienceSciBench chemistry
Score90.61
3
Showing 9 of 9 rows