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

General Task Performance on Macro-average (Mathematics, Multi-Hop QA, Code Generation)

69Accuracy

Claude-Sonnet-4-5

25.73636.96848.259.432Jan 29, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
6991.4
67.588.5
65.476.3
64.687.4
63.981.4
61.982.3
60.276.7
6079.1
59.978.4
2026.01
59.171.5
57.575
56.167.9
2026.01
56.181.1
55.977.7
2026.01
54.864.9
50.565.4
48.475.1
47.770.3
45.972.7
40.570.8
27.464.1