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

Active Reasoning on ARBench

46.7DC Score

OpenAI o3-mini (medium)

35.46838.38441.344.216Apr 17, 2026
Updated 2mo ago

Evaluation Results

MethodLinks
46.743.145.1
2026.04
45.341.543.9
2026.04
44.339.443.5
4440.843.6
2026.04
43.738.743.1
2026.04
43.138.442.3
2026.04
43.13842.3
2026.04
42.738.941.1
42.135.440.7
2026.04
41.936.139.5
2026.04
41.335.739.1
2026.04
40.735.939.3
2026.04
40.135.238.9
2026.04
39.534.138.1
2026.04
37.932.736.5
2026.04
35.932.734.9