1
Claude Opus 4.8, GPT-5.5, and Gemini 2.5 Pro All Score Exactly 50% on Adversarial Safety — The Frontier Ceiling Is Real
五大前沿AI模型在对抗安全测试中齐刷刷得50%，揭示AI安全天花板真实存在。
Every frontier model provider tells you their latest model is "the safest yet." So I tested that claim — on the newest models money can buy. I ran Cla…