ResearchBenchmarksResponse Reliability Estimation on ArabicaQAFollow0.647AUROCGLU0.483720.526110.56850.61089Jun 2, 2026Evaluation ResultsMethodMethodLinksAUROCPRRGLUModel=GemmaModel=Gemma2026.060.6470.322RAUQModel=GemmaModel=Gemma2026.060.6420.277RAUQModel=FanarModel=Fanar2026.060.6130.239RAUQModel=QwenModel=Qwen2026.060.610.186GLUModel=FanarModel=Fanar2026.060.5630.143GLUModel=QwenModel=Qwen2026.060.5560.14LogTokUModel=GemmaModel=Gemma2026.060.5380.099LogTokUModel=QwenModel=Qwen2026.060.5210.028LogTokUModel=FanarModel=Fanar2026.060.490.025