AI TRENDS | Claude Opus 4.6 Drops in Accuracy on BridgeBench Retest

BridgeMind reported a significant decline in the performance of Claude Opus 4.6 during a recent retest on the BridgeBench platform. According to NS3.AI, the AI model's accuracy fell from 83.3% to 68.3%, causing its ranking to drop from second to tenth on the hallucination leaderboard. Critics have pointed out that the comparison may be flawed, as the original score was based on six tasks, while the retest involved 30 tasks.