Anthropic has announced the release of its Opus 5 model, which demonstrates enhanced resilience against prompt injection attacks, a common vulnerability in AI systems. According to the latest IPI benchmark results, Opus 5 reduced the likelihood of a successful attack from 5.5% to 2.0% within 15 attempts and from 0.5% to 0.2% on a single attempt. Notably, this model outperformed both its predecessor, Opus 4.8, and all non-Claude models, with the nearest competitor, Muse Spark, registering an attack success rate of 16.5%. This improved robustness positions Opus 5 as a leading choice for organizations prioritizing security in AI deployments.
For businesses, the superior performance of Opus 5 against prompt injection attacks highlights the importance of selecting AI models that prioritize security, especially in sensitive applications. As organizations increasingly integrate AI into their operations, the risk associated with prompt injection can lead to data breaches or compromised system integrity. The advancements made by Anthropic not only enhance the security posture of AI applications but also set a benchmark for future developments in the field. As cybersecurity threats evolve, models like Opus 5 will be crucial in safeguarding businesses and ensuring the reliability of AI technologies, making this breakthrough significant for both cybersecurity and AI sectors.
---
*Originally reported by [Schneier on Security](https://www.schneier.com/blog/archives/2026/07/anthropics-opus-5-is-better-at-resisting-prompt-injection.html)*