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

Math on Minerva

89.16Accuracy

COSE

Updated 5d ago

Evaluation Results

Method	Links
COSE 2026.05		89.16
COSE 2026.05		88.38
MAE 2026.05		81.8
R-Zero 2026.05		81.2
AZR 2026.05		81.1
Base 2026.05		78.7
MAE 2026.05		70.2
R-Zero 2026.05		69.1
AZR 2026.05		68.4
Base 2026.05		64.1
Lowest Centroid 2026.04		58.1
Lowest Centroid 2026.04		57
Pass@1 2026.04		56.3
Lowest Centroid 2026.04		55.9
Bottom Window 2026.04		55.9
Bottom Window 2026.04		55.9
Greedy Decoding 2026.04		55.5
Tail Confidence 2026.04		55.5
Self-Certainty 2026.04		55.2
Pass@1 2026.04		55
Greedy Decoding 2026.04		54.8
Bottom Window 2026.04		54.8
Tail Confidence 2026.04		54.8
Pass@1 2026.04		54.5
Self-Certainty 2026.04		54.4
Tail Confidence 2026.04		54.4
Self-Certainty 2026.04		54
Greedy Decoding 2026.04		53.3
Lowest Centroid 2026.04		50.4
Greedy Decoding 2026.04		48.2
Self-Certainty 2026.04		48.2
Pass@1 2026.04		46.7
Bottom Window 2026.04		46.3
Tail Confidence 2026.04		46
Lowest Centroid 2026.04		40.1
Pass@1 2026.04		38.6
Tail Confidence 2026.04		38.2
Bottom Window 2026.04		38.2
Greedy Decoding 2026.04		37.9
Self-Certainty 2026.04		36