List your company

Commonsense reasoning on Arabic culture (ArabCulture)

57.2AccuracyGamayun
28.91236.25643.650.944Dec 25, 2025

Evaluation results

MethodLinks
Gamayunscenario=zero-shot2025.1257.2
Qwen2.5scenario=zero-shot2025.1254.6
Qwen3scenario=zero-shot, thinking_mode=no-thinking2025.1240.9
EuroLMscenario=zero-shot2025.1235.5
Gemma3scenario=zero-shot2025.1235
Llama3.2scenario=zero-shot2025.1230