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

Active Reasoning on AR-Bench-DC

61Exact Accuracy

BALAR

2.7617.883348.12May 1, 2026May 2, 2026May 3, 2026May 4, 2026May 5, 2026May 6, 2026
Updated 2mo ago

Evaluation Results

MethodLinks
2026.05
61
2026.05
54
2026.05
49
2026.05
47
2026.05
46.7
2026.05
46.7
2026.05
45.5
2026.05
44.3
2026.05
44
2026.05
43.7
2026.05
43.1
2026.05
43.1
2026.05
43
2026.05
42.7
2026.05
42.1
2026.05
41.9
2026.05
41
2026.05
40.7
2026.05
40.1
2026.05
37
2026.05
15
2026.05
8
2026.05
5