Boris Cherny· @bcherny · X·· 2026-07-25AI 评分54
AI 导读
Anthropic 的 Opus 5 在编码、数据分析、生物学等多项评估中达到新 SOTA。更关键的是，它成为该公司迄今最难被提示词注入的模型。结合强对齐、注入探测与 Claude Code 的 Auto Mode 等多层防御后，提示词注入攻击成功率降至约 0%。
正文 · 原文
Opus 5 is a great model for coding, data analysis, design, biology, knowledge work.
More than any of these eval scores, what is most exciting to me is something else: Opus 5 is our least prompt injectable model yet. It is a bit buried in the system card, but across PI evals and red teaming, Opus 5 is very hard to prompt inject successfully.
And when layering defenses -- strong model alignment, combined with prompt injection probes, combined with Auto Mode in Claude Code -- the success rate for prompt injection attacks drops to ~0. This is new and exciting! More about this soon.
On several coding and knowledge work evaluations, Opus 5 is the new state-of-the-art:在 X 查看被引用的帖子
来源：Boris Cherny · x.com