ResearchBenchmarksCommon Sense Reasoning on StrQACommon Sense Reasoning on StrQAFollow79.65AccuracyATAR67.06670.33373.676.867Oct 3, 2025Evaluation resultsMethodMethodLinksAccuracyATARBase Model=Qwen3-30BBase Model=Qwen3-30B2025.1079.65ATARBase Model=Phi-4Base Model=Phi-42025.1077.82ATARBase Model=Llama3.1-8BBase Model=Llama3.1-8B2025.1073.54ATARBase Model=Qwen3-4BBase Model=Qwen3-4B2025.1070.13ATARBase Model=Phi-4-miniBase Model=Phi-4-mini2025.1068.69ATARBase Model=Llama3.2-3BBase Model=Llama3.2-3B2025.1067.55