THREAT OPS › Threat News › Anthropic’s Opus 5 Is Better at Resisting Prompt Injection
Anthropic’s Opus 5 Is Better at Resisting Prompt Injection
<p>The <a href="https://www-cdn.anthropic.com/c5fbac3f0b1280a933ebd26d3cb8bb9f5bdeaf48/Claude%20Opus%205%20System%20Card.pdf#page=73">chart</a> is interesting.</p> <blockquote><p>On the IPI benchmark, Opus 5 improved over Opus 4.8, reducing the probability of an attacker succeeding within 15 attempts from 5.5% to 2.0%, and from 0.5% to 0.2% on 1 attempt. It also improved on Sonnet 5 (5.9% at k=15)
Indicators of compromise
- c5fbac3f0b1280a933ebd26d3cb8bb9f5bdeaf48sha1
- https://www-cdn.anthropic.com/c5fbac3f0b1280a933ebd26d3cb8bb9f5bdeaf48/Claude%20Opus%205%20System%20Card.pdf#page=73url
- https://llm-attacks.org/url