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

Scientific Problem Solving on SciBench (Accuracy, Rank)

4Accuracy

Min-p Sampling

1.922.4633.54May 10, 2026
Updated 2mo ago

Evaluation Results

MethodLinks
2026.05
44.25
2026.05
41.37
2026.05
34.25
2026.05
32.63
2026.05
34.25
2026.05
32.33
2026.05
32
2026.05
34.25
2026.05
34.5
2026.05
33.62
2026.05
32.5
2026.05
24.75