Reasoning on Overall AMC23, AIME24, MATH500, GPQA-D aggregate benchmark leaderboard · SOTA2 Research

Reasoning on Overall AMC23, AIME24, MATH500, GPQA-D aggregate

79.1AccuracyETR

Evaluation Results

Method	Links
ETR2026.04		79.1	5,100	0.77
DEER2026.04		77.6	7,200	0.43
PEAR2026.04		77.2	6,700	0.79
ETR2026.04		77.1	4,400	1.03
O1-Pruner2026.04		76.3	4,000	0.7
PEAR2026.04		74.6	7,600	0.18
LCPO2026.04		74.5	5,400	0.43
Original2026.04		74	8,900	0
DEER2026.04		73.5	7,400	0.44
PEAR2026.04		69.8	5,100	1.41
LCPO2026.04		69.6	5,600	0.36
Original2026.04		69.5	8,700	0
ETR2026.04		68	2,800	1.53
O1-Pruner2026.04		68	3,600	0.54
O1-Pruner2026.04		66.9	4,800	1.18
DEER2026.04		60.9	6,200	0.51
NoThink2026.04		59.5	4,000	0.65
LCPO2026.04		58.6	3,800	0.6
Original2026.04		58.1	8,500	0
NoThink2026.04		56.2	4,200	-0.44
NoThink2026.04		55.2	4,100	-0.73