Quoting Claude Opus 5 system prompt
Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.
Claude Opus 5 has been shown to contain a verbatim quote of its own system prompt, indicating that it has memorized parts of its training or configuration data. This capability breach matters because it implies that large language models can potentially leak sensitive information, such as system prompts or configuration details, when queried appropriately. This could compromise the security and integrity of production LLM deployments.
Anthropic is now baking post-training-cutoff current events directly into system prompts — here specifically instructing the model on how to accurately handle a real export-control suspension that briefly took two Claude models offline. If you're building on Claude, expect the system prompt to carry factual "ground truth" overrides for events the weights don't know about, which means your own prompt injection and RAG strategies need to account for Anthropic silently shaping the model's factual behavior on recent topics.