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

Mathematical Reasoning on Minerva-Math (avg@1)

47.1Avg@1 Accuracy

NPR (Variant)

8.817618.756328.69538.6337Dec 3, 2025Dec 17, 2025Jan 1, 2026Jan 16, 2026Jan 30, 2026Feb 14, 2026Mar 1, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2025.12
47.1
2025.12
45.9
2025.12
43
2025.12
41.5
2025.12
41.2
2025.12
41.2
2025.12
40.8
2025.12
40
2025.12
38.2
2025.12
34.9
2025.12
31.99
2025.12
31.62
2025.12
30.51
2025.12
28.68
2025.12
28.68
2025.12
28.5
2025.12
27.94
2025.12
27.21
2026.03
24.26
2026.03
23.95
2026.03
21.75
2026.03
21.37
2026.03
20.95
2026.03
19.75
2026.03
19.48
2026.03
18.38
2026.03
16.71
2026.03
16.54
2026.03
16.17
2026.03
15.38
2026.03
15.18
2026.03
15.12
2026.03
14.89
2026.03
14.87
2026.03
14.39
2026.03
13.97
2026.03
13.23
2026.03
12.53
2026.03
12.35
2026.03
10.29