ResearchBenchmarksClaim Verification on ourIR retrievalFollow88Precision (NE)VerifAI81.7683.388586.62Jan 16, 2026Evaluation ResultsMethodMethodLinksPrecision (NE)Precision (S)Precision (C)Precision (wa)Recall (NE)Recall (S)Recall (C)Recall (wa)F1 Score (NE)F1 Score (S)F1 Score (C)F1 Score (wa)AccuracyVerifAIJudge LLM=GPT-4Judge LLM=GPT-42026.0188831185808744838485188483VerifAIJudge LLM=GPT-4oJudge LLM=GPT-4o2026.018290886907922848684128584