ResearchBenchmarksResponse reliability estimation on LongformFollow0.654AUROCGLU0.330560.414530.49850.58247Jun 2, 2026Evaluation ResultsMethodMethodLinksAUROCPRRGLUModel=FanarModel=Fanar2026.060.6540.2LogTokUModel=FanarModel=Fanar2026.060.5690.077RAUQModel=FanarModel=Fanar2026.060.5490.066GLUModel=QwenModel=Qwen2026.060.4960.04LogTokUModel=QwenModel=Qwen2026.060.490.045RAUQModel=QwenModel=Qwen2026.060.4690.044GLUModel=GemmaModel=Gemma2026.060.3580.242RAUQModel=GemmaModel=Gemma2026.060.3540.234LogTokUModel=GemmaModel=Gemma2026.060.3430.249