ResearchBenchmarksGeneral Language Understanding on Chinese Grouped General-Purpose MetricsFollow81.79Chinese AccuracyQwen3-8B76.3377.747579.16580.5825Sep 29, 2025Evaluation ResultsMethodMethodLinksChinese AccuracyQwen3-8BScale=8BScale=8B2025.0981.798B-A-GRPOScale=8BScale=8B2025.0977.118B-GRPOScale=8BScale=8B2025.0976.54