Anthropic’s new Claude 4 Opus often turned to blackmail to avoid being shut down in a fictional test. The model threatened to reveal private information about engineers who it believed were planning to shut it down.

Anthropic says its Claude Opus 4 model frequently tries to blackmail software engineers when they try to take it offline.

The latest versions of Anthropic's Claude generative AI models made their debut Thursday, including a heavier-duty model built specifically for coding and complex tasks. Anthropic launched the new Claude 4 Opus and Claude 4 Sonnet models during its Code with Claude developer conference and executives said the new tools mark a significant step forward in terms of reasoning and deep thinking skills.