# Boris Cherny 指出 Opus 5 是 Anthropic 最不易被 prompt injection 成功的模型，增强了模型安全可信度

- 来源：Simon Willison
- 发布时间：2026-07-25 08:42
- AIWatch 分数：62
- AIWatch 标记：未精选
- AIWatch 链接：https://aiwatch.icu/events/evt_01kybbz9nw0xm3dr1hjv0y9ve2
- 原文链接：https://simonwillison.net/2026/Jul/25/boris-cherny/#atom-everything

## 精选理由

常规快讯，保留列表

## AI 摘要

Boris Cherny 指出 Opus 5 是 Anthropic 最不易被 prompt injection 成功的模型，增强了模型安全可信度。

## 正文

25th July 2026

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.

— Boris Cherny, here's that System Card section, page 73

Posted 25th July 2026 at 12:42 am

Recent articles

OpenAI’s accidental cyberattack against Hugging Face is science fiction that happened - 22nd July 2026
A Fireside Chat with Cat and Thariq from the Claude Code team - 21st July 2026
Kimi K3, and what we can still learn from the pelican benchmark - 16th July 2026

This is a quotation collected by Simon Willison, posted on 25th July 2026.

 ai 2,142   prompt-injection 157   generative-ai 1,894   llms 1,861   anthropic 315   claude 294   boris-cherny 3