List your company

Mathematical Reasoning on Combined Benchmarks AVG

52.03Average AccuracyOGER
12.426822.708432.9943.2716Apr 20, 2026

Evaluation results

MethodLinks
OGERBackbone=Qwen2.5-Math-7B2026.0452.03
LuffyBackbone=Qwen2.5-Math-7B2026.0448.66
GRPOBackbone=Qwen2.5-Math-7B2026.0439.12
OGERBackbone=Qwen2.5-Math-1.5B2026.0436.77
LuffyBackbone=Qwen2.5-Math-1.5B2026.0435.25
GRPOBackbone=Qwen2.5-Math-1.5B2026.0428.69
BaseBackbone=Qwen2.5-Math-7B2026.0428.61
BaseBackbone=Qwen2.5-Math-1.5B2026.0413.95