Results

We tested 31 leading AI models from Anthropic, OpenAI, Google, Meta, and others.

Most didn't know the joke. Many fabricated elaborate stories instead.

Click any card to see what each model said.