Anthropic CEO Dario Amodei submitted a report to the U.S. Congress:

“After our investigation, we found that Kimi K3 may have distilled a model that Anthropic has not yet finished developing.”
A congressman pressed further: “If your model isn’t even built yet, how did it distill it?”
Amodei: “That’s exactly why we’re holding a hearing.”
A new topic has emerged in the AI safety field: Future Model Leakage