List your company

Math Reasoning on Math Level-5

77.8AccuracyTAB
60.1264.7169.373.89Apr 6, 2026

Evaluation results

MethodLinks
TABModel Variant=4B, Budget=10k2026.0477.86,895
TABBudget=10k2026.0477.67,333
TABModel Variant=4B, Budget=8k2026.04775,531
TABBudget=8k2026.0476.85,473
TAB All-SubQBudget=10k2026.0476.86,311
TAB All-SubQBudget=8k2026.0475.85,901
StaticBudget=40962026.0474.611,058
L-JProtocol=Multi-Turn, Strategy=All-SubQ2026.0471.25,744
L-JProtocol=Multi-Turn2026.0470.85,588
TAB All-SubQBudget=5k2026.0470.43,255
StaticBudget=20482026.0470.15,901
L-JProtocol=Multi-Turn, Model Variant=4B2026.0469.54,324
L-JProtocol=Individual2026.0469.35,961
L-JProtocol=Individual, Model Variant=4B2026.0469.34,318
TABModel Variant=4B, Budget=5k2026.0469.13,697
TABBudget=5k2026.0468.13,498
StaticBudget=10242026.0466.63,986
TABModel Variant=4B, Budget=3k2026.0465.62,596
TAB All-SubQBudget=3k2026.0465.42,636
TABBudget=3k2026.04652,516
StaticBudget=5122026.0463.22,660
StaticBudget=2562026.0460.82,013